Elasticsearch Curator Python API¶

The Elasticsearch Curator Python API helps you manage your indices and snapshots.

Note

This documentation is for the Elasticsearch Curator Python API. Documentation for the Elasticsearch Curator CLI – which uses this API and is installed as an entry_point as part of the package – is available in the Elastic guide.

Compatibility¶

The Elasticsearch Curator Python API is compatible with the 5.x Elasticsearch versions, and supports Python versions 2.7 and later.

Example Usage¶

import elasticsearch
import curator

client = elasticsearch.Elasticsearch()

ilo = curator.IndexList(client)
ilo.filter_by_regex(kind='prefix', value='logstash-')
ilo.filter_by_age(source='name', direction='older', timestring='%Y.%m.%d', unit='days', unit_count=30)
delete_indices = curator.DeleteIndices(ilo)
delete_indices.do_action()

Tip

See more examples in the Examples page.

Features¶

The API methods fall into the following categories:

Object Classes build and filter index list or snapshot list objects.
Action Classes act on object classes.
Utilities are helper methods.

Logging¶

The Elasticsearch Curator Python API uses the standard logging library from Python. It inherits two loggers from elasticsearch-py: elasticsearch and elasticsearch.trace. Clients use the elasticsearch logger to log standard activity, depending on the log level. The elasticsearch.trace logger logs requests to the server in JSON format as pretty-printed curl commands that you can execute from the command line. The elasticsearch.trace logger is not inherited from the base logger and must be activated separately.

Contents¶

IndexList¶

class curator.indexlist.IndexList(client)¶

all_indices = None¶: Instance variable. All indices in the cluster at instance creation time. Type: list()

client = None¶: An Elasticsearch Client object Also accessible as an instance variable.

empty_list_check()¶: Raise exception if indices is empty

filter_allocated(key=None, value=None, allocation_type='require', exclude=True)¶

Match indices that have the routing allocation rule of key=value from indices

Parameters:	key – The allocation attribute to check for value – The value to check for allocation_type – Type of allocation to apply exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_by_age(source='name', direction=None, timestring=None, unit=None, unit_count=None, field=None, stats_result='min_value', epoch=None, exclude=False, unit_count_pattern=False)¶

Match indices by relative age calculations.

Parameters:

source – Source of index age. Can be one of ‘name’, ‘creation_date’, or ‘field_stats’
direction – Time to filter, either older or younger
timestring – An strftime string to match the datestamp in an index name. Only used for index filtering by name.
unit – One of seconds, minutes, hours, days, weeks, months, or years.
unit_count – The number of unit (s). unit_count * unit will be calculated out to the relative number of seconds.
unit_count_pattern – A regular expression whose capture group identifies the value for unit_count.
field – A timestamp field name. Only used for field_stats based calculations.
stats_result – Either min_value or max_value. Only used in conjunction with source`=``field_stats` to choose whether to reference the minimum or maximum result value.
epoch – An epoch timestamp used in conjunction with unit and unit_count to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

filter_by_alias(aliases=None, exclude=False)¶

Match indices which are associated with the alias or list of aliases identified by aliases.

An update to Elasticsearch 5.5.0 changes the behavior of this from previous 5.x versions: https://www.elastic.co/guide/en/elasticsearch/reference/5.5/breaking-changes-5.5.html#breaking_55_rest_changes

What this means is that indices must appear in all aliases in list aliases or a 404 error will result, leading to no indices being matched. In older versions, if the index was associated with even one of the aliases in aliases, it would result in a match.

It is unknown if this behavior affects anyone. At the time this was written, no users have been bit by this. The code could be adapted to manually loop if the previous behavior is desired. But if no users complain, this will become the accepted/expected behavior.

Parameters:	aliases (list) – A list of alias names. exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

filter_by_count(count=None, reverse=True, use_age=False, pattern=None, source='creation_date', timestring=None, field=None, stats_result='min_value', exclude=True)¶

Remove indices from the actionable list beyond the number count, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of index is provided–for example, indices matching logstash-%Y.%m.%d–then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older indices.

By setting reverse to False, then index3 will be deleted before index2, which will be deleted before index1

use_age allows ordering indices by age. Age is determined by the index creation date by default, but you can specify an source of name, max_value, or min_value. The name source requires the timestring argument.

Parameters:

count – Filter indices beyond count.
reverse – The filtering direction. (default: True).
use_age – Sort indices by age. source is required in this case.
pattern – Select indices to count from a regular expression pattern. This pattern must have one and only one capture group. This can allow a single count filter instance to operate against any number of matching patterns, and keep count of each index in that group. For example, given a pattern of '^(.*)-\d{6}$', it will match both rollover-000001 and index-999990, but not logstash-2017.10.12. Following the same example, if my cluster also had rollover-000002 through rollover-000010 and index-888888 through index-999999, it will process both groups of indices, and include or exclude the count of each.
source – Source of index age. Can be one of name, creation_date, or field_stats. Default: creation_date
timestring – An strftime string to match the datestamp in an index name. Only used if source name is selected.
field – A timestamp field name. Only used if source field_stats is selected.
stats_result – Either min_value or max_value. Only used if source field_stats is selected. It determines whether to reference the minimum or maximum value of field in each index.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_by_regex(kind=None, value=None, exclude=False)¶

Match indices by regular expression (pattern).

Parameters:

kind – Can be one of: suffix, prefix, regex, or timestring. This option defines what kind of filter you will be building.
value – Depends on kind. It is the strftime string if kind is timestring. It’s used to build the regular expression for other kinds.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

filter_by_shards(number_of_shards=None, shard_filter_behavior='greater_than', exclude=False)¶

Match indices with a given shard count.

Selects all indices with a shard count ‘greater_than’ number_of_shards by default. Use shard_filter_behavior to select indices with shard count ‘greater_than’, ‘greater_than_or_equal’, ‘less_than’, ‘less_than_or_equal’, or ‘equal’ to number_of_shards.

Parameters:	number_of_shards – shard threshold shard_filter_behavior – Do you want to filter on greater_than, greater_than_or_equal, less_than, less_than_or_equal, or equal? exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

filter_by_space(disk_space=None, reverse=True, use_age=False, source='creation_date', timestring=None, field=None, stats_result='min_value', exclude=False, threshold_behavior='greater_than')¶

Remove indices from the actionable list based on space consumed, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of index is provided–for example, indices matching logstash-%Y.%m.%d–then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older indices.

By setting reverse to False, then index3 will be deleted before index2, which will be deleted before index1

use_age allows ordering indices by age. Age is determined by the index creation date by default, but you can specify an source of name, max_value, or min_value. The name source requires the timestring argument.

threshold_behavior, when set to greater_than (default), includes if it the index tests to be larger than disk_space. When set to less_than, it includes if the index is smaller than disk_space

Parameters:

disk_space – Filter indices over n gigabytes
threshold_behavior – Size to filter, either greater_than or less_than. Defaults to greater_than to preserve backwards compatability.
reverse – The filtering direction. (default: True). Ignored if use_age is True
use_age – Sort indices by age. source is required in this case.
source – Source of index age. Can be one of name, creation_date, or field_stats. Default: creation_date
timestring – An strftime string to match the datestamp in an index name. Only used if source name is selected.
field – A timestamp field name. Only used if source field_stats is selected.
stats_result – Either min_value or max_value. Only used if source field_stats is selected. It determines whether to reference the minimum or maximum value of field in each index.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

filter_closed(exclude=True)¶

Filter out closed indices from indices

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_empty(exclude=True)¶

Filter indices with a document count of zero

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_forceMerged(max_num_segments=None, exclude=True)¶

Match any index which has max_num_segments per shard or fewer in the actionable list.

Parameters:	max_num_segments – Cutoff number of segments per shard. exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_ilm(exclude=True)¶

Match indices that have the setting index.lifecycle.name

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_kibana(exclude=True)¶

Match any index named .kibana, .kibana-5, or .kibana-6 in indices. Older releases addressed index names that no longer exist.

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_opened(exclude=True)¶

Filter out opened indices from indices

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

filter_period(period_type='relative', source='name', range_from=None, range_to=None, date_from=None, date_to=None, date_from_format=None, date_to_format=None, timestring=None, unit=None, field=None, stats_result='min_value', intersect=False, week_starts_on='sunday', epoch=None, exclude=False)¶

Match indices with ages within a given period.

Parameters:

period_type – Can be either absolute or relative. Default is relative. date_from and date_to are required when using period_type='absolute'`. ``range_from and range_to are required with ``period_type=’relative’`.
source – Source of index age. Can be one of ‘name’, ‘creation_date’, or ‘field_stats’
range_from – How many unit (s) in the past/future is the origin?
range_to – How many unit (s) in the past/future is the end point?
date_from – The simplified date for the start of the range
date_to – The simplified date for the end of the range. If this value is the same as date_from, the full value of unit will be extrapolated for the range. For example, if unit is months, and date_from and date_to are both 2017.01, then the entire month of January 2017 will be the absolute date range.
date_from_format – The strftime string used to parse date_from
date_to_format – The strftime string used to parse date_to
timestring – An strftime string to match the datestamp in an index name. Only used for index filtering by name.
unit – One of hours, days, weeks, months, or years.
field – A timestamp field name. Only used for field_stats based calculations.
stats_result – Either min_value or max_value. Only used in conjunction with source``=``field_stats to choose whether to reference the minimum or maximum result value.
intersect – Only used when source``=``field_stats. If True, only indices where both min_value and max_value are within the period will be selected. If False, it will use whichever you specified. Default is False to preserve expected behavior.
week_starts_on – Either sunday or monday. Default is sunday
epoch – An epoch timestamp used to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

index_info = None¶: Instance variable. Information extracted from indices, such as segment count, age, etc. Populated at instance creation time, and by other private helper methods, as needed. Type: dict()

indices = None¶: Instance variable. The running list of indices which will be used by an Action class. Populated at instance creation time. Type: list()

iterate_filters(filter_dict)¶

Iterate over the filters defined in config and execute them.

Parameters:	filter_dict – The configuration dictionary

Note

filter_dict should be a dictionary with the following form:

{ 'filters' : [
        {
            'filtertype': 'the_filter_type',
            'key1' : 'value1',
            ...
            'keyN' : 'valueN'
        }
    ]
}

working_list()¶: Return the current value of indices as copy-by-value to prevent list stomping during iterations

SnapshotList¶

class curator.snapshotlist.SnapshotList(client, repository=None)¶

client = None¶: An Elasticsearch Client object. Also accessible as an instance variable.

empty_list_check()¶: Raise exception if snapshots is empty

filter_by_age(source='creation_date', direction=None, timestring=None, unit=None, unit_count=None, epoch=None, exclude=False)¶

Remove snapshots from snapshots by relative age calculations.

Parameters:

source – Source of snapshot age. Can be ‘name’, or ‘creation_date’.
direction – Time to filter, either older or younger
timestring – An strftime string to match the datestamp in an snapshot name. Only used for snapshot filtering by name.
unit – One of seconds, minutes, hours, days, weeks, months, or years.
unit_count – The number of unit (s). unit_count * unit will be calculated out to the relative number of seconds.
epoch – An epoch timestamp used in conjunction with unit and unit_count to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

filter_by_count(count=None, reverse=True, use_age=False, source='creation_date', timestring=None, exclude=True)¶

Remove snapshots from the actionable list beyond the number count, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of snapshot is provided–for example, snapshots matching curator-%Y%m%d%H%M%S– then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older snapshots.

By setting reverse to False, then snapshot3 will be acted on before snapshot2, which will be acted on before snapshot1

use_age allows ordering snapshots by age. Age is determined by the snapshot creation date (as identified by start_time_in_millis) by default, but you can also specify a source of name. The name source requires the timestring argument.

Parameters:

count – Filter snapshots beyond count.
reverse – The filtering direction. (default: True).
use_age – Sort snapshots by age. source is required in this case.
source – Source of snapshot age. Can be one of name, or creation_date. Default: creation_date
timestring – An strftime string to match the datestamp in a snapshot name. Only used if source name is selected.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is True

filter_by_regex(kind=None, value=None, exclude=False)¶

Filter out snapshots not matching the pattern, or in the case of exclude, filter those matching the pattern.

Parameters:

kind – Can be one of: suffix, prefix, regex, or timestring. This option defines what kind of filter you will be building.
value – Depends on kind. It is the strftime string if kind is timestring. It’s used to build the regular expression for other kinds.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

filter_by_state(state=None, exclude=False)¶

Filter out snapshots not matching state, or in the case of exclude, filter those matching state.

Parameters:	state – The snapshot state to filter for. Must be one of `SUCCESS`, `PARTIAL`, `FAILED`, or `IN_PROGRESS`. exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

filter_period(period_type='relative', source='name', range_from=None, range_to=None, date_from=None, date_to=None, date_from_format=None, date_to_format=None, timestring=None, unit=None, week_starts_on='sunday', epoch=None, exclude=False)¶

Match snapshots with ages within a given period.

Parameters:

period_type – Can be either absolute or relative. Default is relative. date_from and date_to are required when using period_type='absolute'`. ``range_from and range_to are required with ``period_type=’relative’`.
source – Source of snapshot age. Can be ‘name’, or ‘creation_date’.
range_from – How many unit (s) in the past/future is the origin?
range_to – How many unit (s) in the past/future is the end point?
date_from – The simplified date for the start of the range
date_to – The simplified date for the end of the range. If this value is the same as date_from, the full value of unit will be extrapolated for the range. For example, if unit is months, and date_from and date_to are both 2017.01, then the entire month of January 2017 will be the absolute date range.
date_from_format – The strftime string used to parse date_from
date_to_format – The strftime string used to parse date_to
timestring – An strftime string to match the datestamp in an snapshot name. Only used for snapshot filtering by name.
unit – One of hours, days, weeks, months, or years.
week_starts_on – Either sunday or monday. Default is sunday
epoch – An epoch timestamp used to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

iterate_filters(config)¶

Iterate over the filters defined in config and execute them.

Parameters:	config – A dictionary of filters, as extracted from the YAML configuration file.

Note

config should be a dictionary with the following form:

{ 'filters' : [
        {
            'filtertype': 'the_filter_type',
            'key1' : 'value1',
            ...
            'keyN' : 'valueN'
        }
    ]
}

most_recent()¶: Return the most recent snapshot based on start_time_in_millis.

repository = None¶: An Elasticsearch repository. Also accessible as an instance variable.

snapshot_info = None¶: Instance variable. Information extracted from snapshots, such as age, etc. Populated by internal method __get_snapshots at instance creation time. Type: dict()

snapshots = None¶: Instance variable. The running list of snapshots which will be used by an Action class. Populated by internal methods __get_snapshots at instance creation time. Type: list()

working_list()¶: Return the current value of snapshots as copy-by-value to prevent list stomping during iterations

Action Classes¶

Alias¶

class curator.actions.Alias(name=None, extra_settings={}, **kwargs)¶

Define the Alias object.

Parameters:	name – The alias name extra_settings (dict, representing the settings.) – Extra settings, including filters and routing. For more information see https://www.elastic.co/guide/en/elasticsearch/reference/current/indices-aliases.html

actions = None¶: The list of actions to perform. Populated by curator.actions.Alias.add and curator.actions.Alias.remove

add(ilo, warn_if_no_indices=False)¶

Create add statements for each index in ilo for alias, then append them to actions. Add any extras that may be there.

Parameters:	ilo – A `curator.indexlist.IndexList` object

body()¶: Return a body string suitable for use with the update_aliases API call.

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Run the API call update_aliases with the results of body()

do_dry_run()¶: Log what the output would be, but take no action.

extra_settings = None¶: Instance variable. Any extra things to add to the alias, like filters, or routing.

name = None¶: Instance variable The strftime parsed version of name.

remove(ilo, warn_if_no_indices=False)¶

Create remove statements for each index in ilo for alias, then append them to actions.

Parameters:	ilo – A `curator.indexlist.IndexList` object

warn_if_no_indices = None¶: Instance variable. Preset default value to False.

Allocation¶

class curator.actions.Allocation(ilo, key=None, value=None, allocation_type='require', wait_for_completion=False, wait_interval=3, max_wait=-1)¶

Parameters:

ilo – A curator.indexlist.IndexList object
key – An arbitrary metadata attribute key. Must match the key assigned to at least some of your nodes to have any effect.
value – An arbitrary metadata attribute value. Must correspond to values associated with key assigned to at least some of your nodes to have any effect. If a None value is provided, it will remove any setting associated with that key.
allocation_type – Type of allocation to apply. Default is require
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: False)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion

Note

See: https://www.elastic.co/guide/en/elasticsearch/reference/current/shard-allocation-filtering.html

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Change allocation settings for indices in index_list.indices with the settings in body.

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

wfc = None¶: Instance variable. Internal reference to wait_for_completion

Close¶

class curator.actions.Close(ilo, delete_aliases=False)¶

Parameters:	ilo – A `curator.indexlist.IndexList` object delete_aliases (bool) – If True, will delete any associated aliases before closing indices.

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

delete_aliases = None¶: Instance variable. Internal reference to delete_aliases

do_action()¶: Close open indices in index_list.indices

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

ClusterRouting¶

class curator.actions.ClusterRouting(client, routing_type=None, setting=None, value=None, wait_for_completion=False, wait_interval=9, max_wait=-1)¶

For now, the cluster routing settings are hardcoded to be transient

Parameters:

client – An elasticsearch.Elasticsearch client object
routing_type – Type of routing to apply. Either allocation or rebalance
setting – Currently, the only acceptable value for setting is enable. This is here in case that changes.
value – Used only if setting is enable. Semi-dependent on routing_type. Acceptable values for allocation and rebalance are all, primaries, and none (string, not NoneType). If routing_type is allocation, this can also be new_primaries, and if rebalance, it can be replicas.
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: False)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion

client = None¶: Instance variable. An elasticsearch.Elasticsearch client object

do_action()¶: Change cluster routing settings with the settings in body.

do_dry_run()¶: Log what the output would be, but take no action.

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

wfc = None¶: Instance variable. Internal reference to wait_for_completion

CreateIndex¶

class curator.actions.CreateIndex(client, name, extra_settings={})¶

Parameters:	client – An `elasticsearch.Elasticsearch` client object name – A name, which can contain `time.strftime()` strings extra_settings (dict, representing the settings and mappings.) – The settings and mappings for the index. For more information see https://www.elastic.co/guide/en/elasticsearch/reference/current/indices-create-index.html

body = None¶: Instance variable. Extracted from the config yaml, it should be a dictionary of mappings and settings suitable for index creation.

client = None¶: Instance variable. An elasticsearch.Elasticsearch client object

do_action()¶: Create index identified by name with settings in body

do_dry_run()¶: Log what the output would be, but take no action.

name = None¶: Instance variable. The parsed version of name

DeleteIndices¶

class curator.actions.DeleteIndices(ilo, master_timeout=30)¶

Parameters:	ilo – A `curator.indexlist.IndexList` object master_timeout – Number of seconds to wait for master node response

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Delete indices in index_list.indices

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

master_timeout = None¶: Instance variable. String value of master_timeout + ‘s’, for seconds.

DeleteSnapshots¶

class curator.actions.DeleteSnapshots(slo, retry_interval=120, retry_count=3)¶

Parameters:	slo – A `curator.snapshotlist.SnapshotList` object retry_interval – Number of seconds to delay betwen retries. Default: 120 (seconds) retry_count – Number of attempts to make. Default: 3

client = None¶: Instance variable. The Elasticsearch Client object derived from slo

do_action()¶: Delete snapshots in slo Retry up to retry_count times, pausing retry_interval seconds between retries.

do_dry_run()¶: Log what the output would be, but take no action.

repository = None¶: Instance variable. The repository name derived from slo

retry_count = None¶: Instance variable. Internally accessible copy of retry_count

retry_interval = None¶: Instance variable. Internally accessible copy of retry_interval

snapshot_list = None¶: Instance variable. Internal reference to slo

ForceMerge¶

class curator.actions.ForceMerge(ilo, max_num_segments=None, delay=0)¶

Parameters:	ilo – A `curator.indexlist.IndexList` object max_num_segments – Number of segments per shard to forceMerge delay – Number of seconds to delay between forceMerge operations

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

delay = None¶: Instance variable. Internally accessible copy of delay

do_action()¶: forcemerge indices in index_list.indices

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

max_num_segments = None¶: Instance variable. Internally accessible copy of max_num_segments

IndexSettings¶

class curator.actions.IndexSettings(ilo, index_settings={}, ignore_unavailable=False, preserve_existing=False)¶

Parameters:

ilo – A curator.indexlist.IndexList object
index_settings – A dictionary structure with one or more index settings to change.
ignore_unavailable – Whether specified concrete indices should be ignored when unavailable (missing or closed)
preserve_existing – Whether to update existing settings. If set to True existing settings on an index remain unchanged. The default is False

body = None¶: Instance variable. Internal reference to index_settings

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_dry_run()¶: Log what the output would be, but take no action.

ignore_unavailable = None¶: Instance variable. Internal reference to ignore_unavailable

index_list = None¶: Instance variable. Internal reference to ilo

preserve_existing = None¶: Instance variable. Internal reference to preserve_settings

Open¶

class curator.actions.Open(ilo)¶

Parameters:	ilo – A `curator.indexlist.IndexList` object

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Open closed indices in index_list.indices

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

Reindex¶

class curator.actions.Reindex(ilo, request_body, refresh=True, requests_per_second=-1, slices=1, timeout=60, wait_for_active_shards=1, wait_for_completion=True, max_wait=-1, wait_interval=9, remote_url_prefix=None, remote_ssl_no_validate=None, remote_certificate=None, remote_client_cert=None, remote_client_key=None, remote_aws_key=None, remote_aws_secret_key=None, remote_aws_region=None, remote_filters={}, migration_prefix='', migration_suffix='')¶

Parameters:

ilo – A curator.indexlist.IndexList object
request_body – The body to send to elasticsearch.Elasticsearch.reindex(), which must be complete and usable, as Curator will do no vetting of the request_body. If it fails to function, Curator will return an exception.
refresh (bool) – Whether to refresh the entire target index after the operation is complete. (default: True)
requests_per_second – The throttle to set on this request in sub-requests per second. -1 means set no throttle as does unlimited which is the only non-float this accepts. (default: -1)
slices – The number of slices this task should be divided into. 1 means the task will not be sliced into subtasks. (default: 1)
timeout – The length in seconds each individual bulk request should wait for shards that are unavailable. (default: 60)
wait_for_active_shards – Sets the number of shard copies that must be active before proceeding with the reindex operation. (default: 1) means the primary shard only. Set to all for all shard copies, otherwise set to any non-negative value less than or equal to the total number of copies for the shard (number of replicas + 1)
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: True)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion
remote_url_prefix (str) – Optional url prefix, if needed to reach the Elasticsearch API (i.e., it’s not at the root level)
remote_ssl_no_validate (bool) – If True, do not validate the certificate chain. This is an insecure option and you will see warnings in the log output.
remote_certificate – Path to SSL/TLS certificate
remote_client_cert – Path to SSL/TLS client certificate (public key)
remote_client_key – Path to SSL/TLS private key
remote_aws_key – AWS IAM Access Key (Only used if the requests-aws4auth python module is installed)
remote_aws_secret_key – AWS IAM Secret Access Key (Only used if the requests-aws4auth python module is installed)
remote_aws_region – AWS Region (Only used if the requests-aws4auth python module is installed)
remote_filters – Apply these filters to the remote client for remote index selection.
migration_prefix – When migrating, prepend this value to the index name.
migration_suffix – When migrating, append this value to the index name.

body = None¶: Instance variable. Internal reference to request_body

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Execute elasticsearch.Elasticsearch.reindex() operation with the provided request_body and arguments.

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

mpfx = None¶: Instance variable. Internal reference to migration_prefix

msfx = None¶: Instance variable. Internal reference to migration_suffix

refresh = None¶: Instance variable. Internal reference to refresh

requests_per_second = None¶: Instance variable. Internal reference to requests_per_second

show_run_args(source, dest)¶: Show what will run

slices = None¶: Instance variable. Internal reference to slices

timeout = None¶: Instance variable. Internal reference to timeout, and add “s” for seconds.

wait_for_active_shards = None¶: Instance variable. Internal reference to wait_for_active_shards

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

wfc = None¶: Instance variable. Internal reference to wait_for_completion

Replicas¶

class curator.actions.Replicas(ilo, count=None, wait_for_completion=False, wait_interval=9, max_wait=-1)¶

Parameters:	ilo – A `curator.indexlist.IndexList` object count – The count of replicas per shard wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: False) wait_interval – How long in seconds to wait between checks for completion. max_wait – Maximum number of seconds to wait_for_completion

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

count = None¶: Instance variable. Internally accessible copy of count

do_action()¶: Update the replica count of indices in index_list.indices

do_dry_run()¶: Log what the output would be, but take no action.

index_list = None¶: Instance variable. Internal reference to ilo

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

wfc = None¶: Instance variable. Internal reference to wait_for_completion

Restore¶

class curator.actions.Restore(slo, name=None, indices=None, include_aliases=False, ignore_unavailable=False, include_global_state=False, partial=False, rename_pattern=None, rename_replacement=None, extra_settings={}, wait_for_completion=True, wait_interval=9, max_wait=-1, skip_repo_fs_check=False)¶

Parameters:

slo – A curator.snapshotlist.SnapshotList object
name (str) – Name of the snapshot to restore. If no name is provided, it will restore the most recent snapshot by age.
indices (list) – A list of indices to restore. If no indices are provided, it will restore all indices in the snapshot.
include_aliases (bool) – If set to True, restore aliases with the indices. (default: False)
ignore_unavailable (bool) – Ignore unavailable shards/indices. (default: False)
include_global_state (bool) – Restore cluster global state with snapshot. (default: False)
partial (bool) – Do not fail if primary shard is unavailable. (default: False)
rename_pattern (str) – A regular expression pattern with one or more captures, e.g. index_(.+)
rename_replacement (str) – A target index name pattern with $# numbered references to the captures in rename_pattern, e.g. restored_index_$1
extra_settings (dict, representing the settings.) – Extra settings, including shard count and settings to omit. For more information see https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-snapshots.html#_changing_index_settings_during_restore
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: True)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion
skip_repo_fs_check (bool) – Do not validate write access to repository on all cluster nodes before proceeding. (default: False). Useful for shared filesystems where intermittent timeouts can affect validation, but won’t likely affect snapshot success.

body = None¶: Instance variable. Populated at instance creation time from the other options

client = None¶: Instance variable. The Elasticsearch Client object derived from slo

do_action()¶: Restore indices with options passed.

do_dry_run()¶: Log what the output would be, but take no action.

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

name = None¶: Instance variable. Will use a provided snapshot name, or the most recent snapshot in slo

py_rename_replacement = None¶: Also an instance variable version of rename_replacement but with Java regex group designations of $# converted to Python’s \\# style.

rename_pattern = None¶: Instance variable version of rename_pattern

rename_replacement = None¶: Instance variable version of rename_replacement

report_state()¶: Log the state of the restore This should only be done if wait_for_completion is True, and only after completing the restore.

repository = None¶: Instance variable. repository derived from slo

skip_repo_fs_check = None¶: Instance variable. Internally accessible copy of skip_repo_fs_check

snapshot_list = None¶: Instance variable. Internal reference to slo

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

Rollover¶

class curator.actions.Rollover(client, name, conditions, new_index=None, extra_settings=None, wait_for_active_shards=1)¶

Parameters:

client – An elasticsearch.Elasticsearch client object
name – The name of the single-index-mapped alias to test for rollover conditions.
conditions – A dictionary of conditions to test
extra_settings – Must be either None, or a dictionary of settings to apply to the new index on rollover. This is used in place of settings in the Rollover API, mostly because it’s already existent in other places here in Curator
wait_for_active_shards – The number of shards expected to be active before returning.

New_index:

The new index name

body()¶: Create a body from conditions and settings

client = None¶: Instance variable. The Elasticsearch Client object

conditions = None¶: Instance variable. Internal reference to conditions

do_action()¶: Rollover the index referenced by alias name

do_dry_run()¶: Log what the output would be, but take no action.

doit(dry_run=False)¶: This exists solely to prevent having to have duplicate code in both do_dry_run and do_action

new_index = None¶: Instance variable. Internal reference to new_index

settings = None¶: Instance variable. Internal reference to extra_settings

wait_for_active_shards = None¶: Instance variable. Internal reference to wait_for_active_shards

Shrink¶

class curator.actions.Shrink(ilo, shrink_node='DETERMINISTIC', node_filters={}, number_of_shards=1, number_of_replicas=1, shrink_prefix='', shrink_suffix='-shrink', copy_aliases=False, delete_after=True, post_allocation={}, wait_for_active_shards=1, wait_for_rebalance=True, extra_settings={}, wait_for_completion=True, wait_interval=9, max_wait=-1)¶

Parameters:

ilo – A curator.indexlist.IndexList object
shrink_node – The node name to use as the shrink target, or DETERMINISTIC, which will use the values in node_filters to determine which node will be the shrink node.
node_filters (dict, representing the filters) – If the value of shrink_node is DETERMINISTIC, the values in node_filters will be used while determining which node to allocate the shards on before performing the shrink.
number_of_shards – The number of shards the shrunk index should have
number_of_replicas – The number of replicas for the shrunk index
shrink_prefix – Prepend the shrunk index with this value
shrink_suffix – Append the value to the shrunk index (default: -shrink)
copy_aliases (bool) – Whether to copy each source index aliases to target index after shrinking. the aliases will be added to target index and deleted from source index at the same time(default: False)
delete_after (bool) – Whether to delete each index after shrinking. (default: True)
post_allocation (dict, with keys allocation_type, key, and value) – If populated, the allocation_type, key, and value will be applied to the shrunk index to re-route it.
wait_for_active_shards – The number of shards expected to be active before returning.
extra_settings (dict) – Permitted root keys are settings and aliases. See https://www.elastic.co/guide/en/elasticsearch/reference/current/indices-shrink-index.html
wait_for_rebalance (bool) – Wait for rebalance. (default: True)
wait_for_active_shards – Wait for active shards before returning.
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. You should not normally change this, ever. (default: True)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

copy_aliases = None¶: Instance variable. Internal reference to copy_aliases

delete_after = None¶: Instance variable. Internal reference to delete_after

do_dry_run()¶: Show what a regular run would do, but don’t actually do it.

index_list = None¶: Instance variable. Internal reference to ilo

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

most_available_node()¶

Determine which data node name has the most available free space, and meets the other node filters settings.

Parameters:	client – An `elasticsearch.Elasticsearch` client object

node_filters = None¶: Instance variable. Internal reference to node_filters

number_of_shards = None¶: Instance variable. Internal reference to number_of_shards

post_allocation = None¶: Instance variable. Internal reference to post_allocation

shrink_node = None¶: Instance variable. Internal reference to shrink_node

shrink_prefix = None¶: Instance variable. Internal reference to shrink_prefix

shrink_suffix = None¶: Instance variable. Internal reference to shrink_suffix

wait_for_rebalance = None¶: Instance variable. Internal reference to wait_for_rebalance

wait_interval = None¶: Instance variable. How many seconds to wait between checks for completion.

wfc = None¶: Instance variable. Internal reference to wait_for_completion

Snapshot¶

class curator.actions.Snapshot(ilo, repository=None, name=None, ignore_unavailable=False, include_global_state=True, partial=False, wait_for_completion=True, wait_interval=9, max_wait=-1, skip_repo_fs_check=False)¶

Parameters:

ilo – A curator.indexlist.IndexList object
repository – The Elasticsearch snapshot repository to use
name – What to name the snapshot.
wait_for_completion (bool) – Wait (or not) for the operation to complete before returning. (default: True)
wait_interval – How long in seconds to wait between checks for completion.
max_wait – Maximum number of seconds to wait_for_completion
ignore_unavailable (bool) – Ignore unavailable shards/indices. (default: False)
include_global_state (bool) – Store cluster global state with snapshot. (default: True)
partial (bool) – Do not fail if primary shard is unavailable. (default: False)
skip_repo_fs_check (bool) – Do not validate write access to repository on all cluster nodes before proceeding. (default: False). Useful for shared filesystems where intermittent timeouts can affect validation, but won’t likely affect snapshot success.

body = None¶: Instance variable. Populated at instance creation time by calling curator.utils.utils.create_snapshot_body with ilo.indices and the provided arguments: ignore_unavailable, include_global_state, partial

client = None¶: Instance variable. The Elasticsearch Client object derived from ilo

do_action()¶: Snapshot indices in index_list.indices, with options passed.

do_dry_run()¶: Log what the output would be, but take no action.

get_state()¶: Get the state of the snapshot

index_list = None¶: Instance variable. Internal reference to ilo

max_wait = None¶: Instance variable. How long in seconds to wait_for_completion before returning with an exception. A value of -1 means wait forever.

name = None¶: Instance variable. The parsed version of name

report_state()¶: Log the state of the snapshot and raise an exception if the state is not SUCCESS

repository = None¶: Instance variable. Internally accessible copy of repository

skip_repo_fs_check = None¶: Instance variable. Internally accessible copy of skip_repo_fs_check

wait_for_completion = None¶: Instance variable. Internally accessible copy of wait_for_completion

wait_interval = None¶: Instance variable How many seconds to wait between checks for completion.

Filter Methods¶

IndexList
SnapshotList

IndexList¶

IndexList.filter_allocated(key=None, value=None, allocation_type='require', exclude=True)

Match indices that have the routing allocation rule of key=value from indices

Parameters:	key – The allocation attribute to check for value – The value to check for allocation_type – Type of allocation to apply exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_by_age(source='name', direction=None, timestring=None, unit=None, unit_count=None, field=None, stats_result='min_value', epoch=None, exclude=False, unit_count_pattern=False)

Match indices by relative age calculations.

Parameters:

source – Source of index age. Can be one of ‘name’, ‘creation_date’, or ‘field_stats’
direction – Time to filter, either older or younger
timestring – An strftime string to match the datestamp in an index name. Only used for index filtering by name.
unit – One of seconds, minutes, hours, days, weeks, months, or years.
unit_count – The number of unit (s). unit_count * unit will be calculated out to the relative number of seconds.
unit_count_pattern – A regular expression whose capture group identifies the value for unit_count.
field – A timestamp field name. Only used for field_stats based calculations.
stats_result – Either min_value or max_value. Only used in conjunction with source`=``field_stats` to choose whether to reference the minimum or maximum result value.
epoch – An epoch timestamp used in conjunction with unit and unit_count to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

IndexList.filter_by_regex(kind=None, value=None, exclude=False)

Match indices by regular expression (pattern).

Parameters:

kind – Can be one of: suffix, prefix, regex, or timestring. This option defines what kind of filter you will be building.
value – Depends on kind. It is the strftime string if kind is timestring. It’s used to build the regular expression for other kinds.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

IndexList.filter_by_space(disk_space=None, reverse=True, use_age=False, source='creation_date', timestring=None, field=None, stats_result='min_value', exclude=False, threshold_behavior='greater_than')

Remove indices from the actionable list based on space consumed, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of index is provided–for example, indices matching logstash-%Y.%m.%d–then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older indices.

By setting reverse to False, then index3 will be deleted before index2, which will be deleted before index1

use_age allows ordering indices by age. Age is determined by the index creation date by default, but you can specify an source of name, max_value, or min_value. The name source requires the timestring argument.

threshold_behavior, when set to greater_than (default), includes if it the index tests to be larger than disk_space. When set to less_than, it includes if the index is smaller than disk_space

Parameters:

disk_space – Filter indices over n gigabytes
threshold_behavior – Size to filter, either greater_than or less_than. Defaults to greater_than to preserve backwards compatability.
reverse – The filtering direction. (default: True). Ignored if use_age is True
use_age – Sort indices by age. source is required in this case.
source – Source of index age. Can be one of name, creation_date, or field_stats. Default: creation_date
timestring – An strftime string to match the datestamp in an index name. Only used if source name is selected.
field – A timestamp field name. Only used if source field_stats is selected.
stats_result – Either min_value or max_value. Only used if source field_stats is selected. It determines whether to reference the minimum or maximum value of field in each index.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

IndexList.filter_closed(exclude=True)

Filter out closed indices from indices

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_forceMerged(max_num_segments=None, exclude=True)

Match any index which has max_num_segments per shard or fewer in the actionable list.

Parameters:	max_num_segments – Cutoff number of segments per shard. exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_kibana(exclude=True)

Match any index named .kibana, .kibana-5, or .kibana-6 in indices. Older releases addressed index names that no longer exist.

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_opened(exclude=True)

Filter out opened indices from indices

Parameters:	exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_none()

IndexList.filter_by_alias(aliases=None, exclude=False)

Match indices which are associated with the alias or list of aliases identified by aliases.

An update to Elasticsearch 5.5.0 changes the behavior of this from previous 5.x versions: https://www.elastic.co/guide/en/elasticsearch/reference/5.5/breaking-changes-5.5.html#breaking_55_rest_changes

What this means is that indices must appear in all aliases in list aliases or a 404 error will result, leading to no indices being matched. In older versions, if the index was associated with even one of the aliases in aliases, it would result in a match.

It is unknown if this behavior affects anyone. At the time this was written, no users have been bit by this. The code could be adapted to manually loop if the previous behavior is desired. But if no users complain, this will become the accepted/expected behavior.

Parameters:	aliases (list) – A list of alias names. exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

IndexList.filter_by_count(count=None, reverse=True, use_age=False, pattern=None, source='creation_date', timestring=None, field=None, stats_result='min_value', exclude=True)

Remove indices from the actionable list beyond the number count, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of index is provided–for example, indices matching logstash-%Y.%m.%d–then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older indices.

By setting reverse to False, then index3 will be deleted before index2, which will be deleted before index1

use_age allows ordering indices by age. Age is determined by the index creation date by default, but you can specify an source of name, max_value, or min_value. The name source requires the timestring argument.

Parameters:

count – Filter indices beyond count.
reverse – The filtering direction. (default: True).
use_age – Sort indices by age. source is required in this case.
pattern – Select indices to count from a regular expression pattern. This pattern must have one and only one capture group. This can allow a single count filter instance to operate against any number of matching patterns, and keep count of each index in that group. For example, given a pattern of '^(.*)-\d{6}$', it will match both rollover-000001 and index-999990, but not logstash-2017.10.12. Following the same example, if my cluster also had rollover-000002 through rollover-000010 and index-888888 through index-999999, it will process both groups of indices, and include or exclude the count of each.
source – Source of index age. Can be one of name, creation_date, or field_stats. Default: creation_date
timestring – An strftime string to match the datestamp in an index name. Only used if source name is selected.
field – A timestamp field name. Only used if source field_stats is selected.
stats_result – Either min_value or max_value. Only used if source field_stats is selected. It determines whether to reference the minimum or maximum value of field in each index.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is True

IndexList.filter_period(period_type='relative', source='name', range_from=None, range_to=None, date_from=None, date_to=None, date_from_format=None, date_to_format=None, timestring=None, unit=None, field=None, stats_result='min_value', intersect=False, week_starts_on='sunday', epoch=None, exclude=False)

Match indices with ages within a given period.

Parameters:

period_type – Can be either absolute or relative. Default is relative. date_from and date_to are required when using period_type='absolute'`. ``range_from and range_to are required with ``period_type=’relative’`.
source – Source of index age. Can be one of ‘name’, ‘creation_date’, or ‘field_stats’
range_from – How many unit (s) in the past/future is the origin?
range_to – How many unit (s) in the past/future is the end point?
date_from – The simplified date for the start of the range
date_to – The simplified date for the end of the range. If this value is the same as date_from, the full value of unit will be extrapolated for the range. For example, if unit is months, and date_from and date_to are both 2017.01, then the entire month of January 2017 will be the absolute date range.
date_from_format – The strftime string used to parse date_from
date_to_format – The strftime string used to parse date_to
timestring – An strftime string to match the datestamp in an index name. Only used for index filtering by name.
unit – One of hours, days, weeks, months, or years.
field – A timestamp field name. Only used for field_stats based calculations.
stats_result – Either min_value or max_value. Only used in conjunction with source``=``field_stats to choose whether to reference the minimum or maximum result value.
intersect – Only used when source``=``field_stats. If True, only indices where both min_value and max_value are within the period will be selected. If False, it will use whichever you specified. Default is False to preserve expected behavior.
week_starts_on – Either sunday or monday. Default is sunday
epoch – An epoch timestamp used to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

SnapshotList¶

SnapshotList.filter_by_age(source='creation_date', direction=None, timestring=None, unit=None, unit_count=None, epoch=None, exclude=False)

Remove snapshots from snapshots by relative age calculations.

Parameters:

source – Source of snapshot age. Can be ‘name’, or ‘creation_date’.
direction – Time to filter, either older or younger
timestring – An strftime string to match the datestamp in an snapshot name. Only used for snapshot filtering by name.
unit – One of seconds, minutes, hours, days, weeks, months, or years.
unit_count – The number of unit (s). unit_count * unit will be calculated out to the relative number of seconds.
epoch – An epoch timestamp used in conjunction with unit and unit_count to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

SnapshotList.filter_by_regex(kind=None, value=None, exclude=False)

Filter out snapshots not matching the pattern, or in the case of exclude, filter those matching the pattern.

Parameters:

kind – Can be one of: suffix, prefix, regex, or timestring. This option defines what kind of filter you will be building.
value – Depends on kind. It is the strftime string if kind is timestring. It’s used to build the regular expression for other kinds.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

SnapshotList.filter_by_state(state=None, exclude=False)

Filter out snapshots not matching state, or in the case of exclude, filter those matching state.

Parameters:	state – The snapshot state to filter for. Must be one of `SUCCESS`, `PARTIAL`, `FAILED`, or `IN_PROGRESS`. exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is False

SnapshotList.filter_none()

SnapshotList.filter_by_count(count=None, reverse=True, use_age=False, source='creation_date', timestring=None, exclude=True)

Remove snapshots from the actionable list beyond the number count, sorted reverse-alphabetically by default. If you set reverse to False, it will be sorted alphabetically.

The default is usually what you will want. If only one kind of snapshot is provided–for example, snapshots matching curator-%Y%m%d%H%M%S– then reverse alphabetical sorting will mean the oldest will remain in the list, because lower numbers in the dates mean older snapshots.

By setting reverse to False, then snapshot3 will be acted on before snapshot2, which will be acted on before snapshot1

use_age allows ordering snapshots by age. Age is determined by the snapshot creation date (as identified by start_time_in_millis) by default, but you can also specify a source of name. The name source requires the timestring argument.

Parameters:

count – Filter snapshots beyond count.
reverse – The filtering direction. (default: True).
use_age – Sort snapshots by age. source is required in this case.
source – Source of snapshot age. Can be one of name, or creation_date. Default: creation_date
timestring – An strftime string to match the datestamp in a snapshot name. Only used if source name is selected.
exclude – If exclude is True, this filter will remove matching snapshots from snapshots. If exclude is False, then only matching snapshots will be kept in snapshots. Default is True

SnapshotList.filter_period(period_type='relative', source='name', range_from=None, range_to=None, date_from=None, date_to=None, date_from_format=None, date_to_format=None, timestring=None, unit=None, week_starts_on='sunday', epoch=None, exclude=False)

Match snapshots with ages within a given period.

Parameters:

period_type – Can be either absolute or relative. Default is relative. date_from and date_to are required when using period_type='absolute'`. ``range_from and range_to are required with ``period_type=’relative’`.
source – Source of snapshot age. Can be ‘name’, or ‘creation_date’.
range_from – How many unit (s) in the past/future is the origin?
range_to – How many unit (s) in the past/future is the end point?
date_from – The simplified date for the start of the range
date_to – The simplified date for the end of the range. If this value is the same as date_from, the full value of unit will be extrapolated for the range. For example, if unit is months, and date_from and date_to are both 2017.01, then the entire month of January 2017 will be the absolute date range.
date_from_format – The strftime string used to parse date_from
date_to_format – The strftime string used to parse date_to
timestring – An strftime string to match the datestamp in an snapshot name. Only used for snapshot filtering by name.
unit – One of hours, days, weeks, months, or years.
week_starts_on – Either sunday or monday. Default is sunday
epoch – An epoch timestamp used to establish a point of reference for calculations. If not provided, the current time will be used.
exclude – If exclude is True, this filter will remove matching indices from indices. If exclude is False, then only matching indices will be kept in indices. Default is False

Utility & Helper Methods¶

class curator.utils.TimestringSearch(timestring)¶

An object to allow repetitive search against a string, searchme, without having to repeatedly recreate the regex.

Parameters:	timestring – An strftime pattern

get_epoch(searchme)¶

Return the epoch timestamp extracted from the timestring appearing in searchme.

Parameters:	searchme – A string to be searched for a date pattern that matches timestring
Return type:	int

curator.utils.absolute_date_range(unit, date_from, date_to, date_from_format=None, date_to_format=None)¶

Get the epoch start time and end time of a range of unit``s, reckoning the start of the week (if that's the selected unit) based on ``week_starts_on, which can be either sunday or monday.

Parameters:

unit – One of hours, days, weeks, months, or years.
date_from – The simplified date for the start of the range
date_to – The simplified date for the end of the range. If this value is the same as date_from, the full value of unit will be extrapolated for the range. For example, if unit is months, and date_from and date_to are both 2017.01, then the entire month of January 2017 will be the absolute date range.
date_from_format – The strftime string used to parse date_from
date_to_format – The strftime string used to parse date_to

Return type:

tuple

curator.utils.byte_size(num, suffix='B')¶

Return a formatted string indicating the size in bytes, with the proper unit, e.g. KB, MB, GB, TB, etc.

Parameters:	num – The number of byte suffix – An arbitrary suffix, like Bytes
Return type:	float

curator.utils.check_csv(value)¶

Some of the curator methods should not operate against multiple indices at once. This method can be used to check if a list or csv has been sent.

Parameters:	value – The value to test, if list or csv string
Return type:	bool

curator.utils.check_master(client, master_only=False)¶

Check if connected client is the elected master node of the cluster. If not, cleanly exit with a log message.

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	None

curator.utils.check_version(client)¶

Verify version is within acceptable range. Raise an exception if it is not.

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	None

curator.utils.chunk_index_list(indices)¶

This utility chunks very large index lists into 3KB chunks It measures the size as a csv string, then converts back into a list for the return value.

Parameters:	indices – A list of indices to act on.
Return type:	list

curator.utils.create_repo_body(repo_type=None, compress=True, chunk_size=None, max_restore_bytes_per_sec=None, max_snapshot_bytes_per_sec=None, location=None, bucket=None, region=None, base_path=None, access_key=None, secret_key=None, **kwargs)¶

Build the ‘body’ portion for use in creating a repository.

Parameters:	repo_type – The type of repository (presently only fs and s3) compress – Turn on compression of the snapshot files. Compression is applied only to metadata files (index mapping and settings). Data files are not compressed. (Default: True) chunk_size – The chunk size can be specified in bytes or by using size value notation, i.e. 1g, 10m, 5k. Defaults to null (unlimited chunk size). max_restore_bytes_per_sec – Throttles per node restore rate. Defaults to `20mb` per second. max_snapshot_bytes_per_sec – Throttles per node snapshot rate. Defaults to `20mb` per second. location – Location of the snapshots. Required. bucket – S3 only. The name of the bucket to be used for snapshots. Required. region – S3 only. The region where bucket is located. Defaults to US Standard base_path – S3 only. Specifies the path within bucket to repository data. Defaults to value of `repositories.s3.base_path` or to root directory if not set. access_key – S3 only. The access key to use for authentication. Defaults to value of `cloud.aws.access_key`. secret_key – S3 only. The secret key to use for authentication. Defaults to value of `cloud.aws.secret_key`.
Returns:	A dictionary suitable for creating a repository from the provided arguments.
Return type:	dict

curator.utils.create_repository(client, **kwargs)¶

Create repository with repository and body settings

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use repo_type – The type of repository (presently only fs and s3) compress – Turn on compression of the snapshot files. Compression is applied only to metadata files (index mapping and settings). Data files are not compressed. (Default: True) chunk_size – The chunk size can be specified in bytes or by using size value notation, i.e. 1g, 10m, 5k. Defaults to null (unlimited chunk size). max_restore_bytes_per_sec – Throttles per node restore rate. Defaults to `20mb` per second. max_snapshot_bytes_per_sec – Throttles per node snapshot rate. Defaults to `20mb` per second. location – Location of the snapshots. Required. bucket – S3 only. The name of the bucket to be used for snapshots. Required. region – S3 only. The region where bucket is located. Defaults to US Standard base_path – S3 only. Specifies the path within bucket to repository data. Defaults to value of `repositories.s3.base_path` or to root directory if not set. access_key – S3 only. The access key to use for authentication. Defaults to value of `cloud.aws.access_key`. secret_key – S3 only. The secret key to use for authentication. Defaults to value of `cloud.aws.secret_key`. skip_repo_fs_check – Skip verifying the repo after creation.
Returns:	A boolean value indicating success or failure.
Return type:	bool

curator.utils.create_snapshot_body(indices, ignore_unavailable=False, include_global_state=True, partial=False)¶

Create the request body for creating a snapshot from the provided arguments.

Parameters:	indices – A single index, or list of indices to snapshot. ignore_unavailable (bool) – Ignore unavailable shards/indices. (default: False) include_global_state (bool) – Store cluster global state with snapshot. (default: True) partial (bool) – Do not fail if primary shard is unavailable. (default: False)
Return type:	dict

curator.utils.date_range(unit, range_from, range_to, epoch=None, week_starts_on='sunday')¶

Get the epoch start time and end time of a range of unit``s, reckoning the start of the week (if that's the selected unit) based on ``week_starts_on, which can be either sunday or monday.

Parameters:	unit – One of `hours`, `days`, `weeks`, `months`, or `years`. range_from – How many `unit` (s) in the past/future is the origin? range_to – How many `unit` (s) in the past/future is the end point? epoch – An epoch timestamp used to establish a point of reference for calculations. week_starts_on – Either `sunday` or `monday`. Default is `sunday`
Return type:	tuple

curator.utils.ensure_list(indices)¶

Return a list, even if indices is a single value

Parameters:	indices – A list of indices to act upon
Return type:	list

curator.utils.find_snapshot_tasks(client)¶

Check if there is snapshot activity in the Tasks API. Return True if activity is found, or False

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	bool

curator.utils.fix_epoch(epoch)¶

Fix value of epoch to be epoch, which should be 10 or fewer digits long.

Parameters:	epoch – An epoch timestamp, in epoch + milliseconds, or microsecond, or even nanoseconds.
Return type:	int

curator.utils.get_client(**kwargs)¶

NOTE: AWS IAM parameters aws_sign_request and aws_region are: provided to facilitate request signing. The credentials will be fetched from the local environment as per the AWS documentation: http://amzn.to/2fRCGCt

AWS IAM parameters aws_key, aws_secret_key, and aws_region are provided for users that still have their keys included in the Curator config file.

Return an elasticsearch.Elasticsearch client object using the provided parameters. Any of the keyword arguments the elasticsearch.Elasticsearch client object can receive are valid, such as:

Parameters:

hosts (list) – A list of one or more Elasticsearch client hostnames or IP addresses to connect to. Can send a single host.
port (int) – The Elasticsearch client port to connect to.
url_prefix (str) – Optional url prefix, if needed to reach the Elasticsearch API (i.e., it’s not at the root level)
use_ssl (bool) – Whether to connect to the client via SSL/TLS
certificate – Path to SSL/TLS certificate
client_cert – Path to SSL/TLS client certificate (public key)
client_key – Path to SSL/TLS private key
aws_key – AWS IAM Access Key (Only used if the requests-aws4auth python module is installed)
aws_secret_key – AWS IAM Secret Access Key (Only used if the requests-aws4auth python module is installed)
aws_region – AWS Region (Only used if the requests-aws4auth python module is installed)
aws_sign_request –

Sign request to AWS (Only used if the requests-aws4auth

and boto3 python modules are installed)

arg aws_region: AWS Region where the cluster exists (Only used if the requests-aws4auth and boto3 python modules are installed)
ssl_no_validate (bool) – If True, do not validate the certificate chain. This is an insecure option and you will see warnings in the log output.
http_auth (str) – Authentication credentials in user:pass format.
timeout (int) – Number of seconds before the client will timeout.
master_only (bool) – If True, the client will only connect if the endpoint is the elected master node of the cluster. This option does not work if `hosts` has more than one value. It will raise an Exception in that case.
skip_version_test – If True, skip the version check as part of the client connection.

Return type:

elasticsearch.Elasticsearch

curator.utils.get_date_regex(timestring)¶

Return a regex string based on a provided strftime timestring.

Parameters:	timestring – An strftime pattern
Return type:	str

curator.utils.get_datemath(client, datemath, random_element=None)¶: Return the parsed index name from datemath

curator.utils.get_datetime(index_timestamp, timestring)¶

Return the datetime extracted from the index name, which is the index creation time.

Parameters:	index_timestamp – The timestamp extracted from an index name timestring – An strftime pattern
Return type:	`datetime.datetime`

curator.utils.get_indices(client)¶

Get the current list of indices from the cluster.

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	list

curator.utils.get_point_of_reference(unit, count, epoch=None)¶

Get a point-of-reference timestamp in epoch + milliseconds by deriving from a unit and a count, and an optional reference timestamp, epoch

Parameters:	unit – One of `seconds`, `minutes`, `hours`, `days`, `weeks`, `months`, or `years`. unit_count – The number of `units`. `unit_count` * `unit` will be calculated out to the relative number of seconds. epoch – An epoch timestamp used in conjunction with `unit` and `unit_count` to establish a point of reference for calculations.
Return type:	int

curator.utils.get_repository(client, repository='')¶

Return configuration information for the indicated repository.

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use
Return type:	dict

curator.utils.get_snapshot(client, repository=None, snapshot='')¶

Return information about a snapshot (or a comma-separated list of snapshots) If no snapshot specified, it will return all snapshots. If none exist, an empty dictionary will be returned.

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use snapshot – The snapshot name, or a comma-separated list of snapshots
Return type:	dict

curator.utils.get_snapshot_data(client, repository=None)¶

Get _all snapshots from repository and return a list.

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use
Return type:	list

curator.utils.get_version(client)¶

Return the ES version number as a tuple. Omits trailing tags like -dev, or Beta

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	tuple

curator.utils.get_yaml(path)¶

Read the file identified by path and import its YAML contents.

Parameters:	path – The path to a YAML configuration file.
Return type:	dict

curator.utils.health_check(client, **kwargs)¶

This function calls client.cluster.health and, based on the args provided, will return True or False depending on whether that particular keyword appears in the output, and has the expected value. If multiple keys are provided, all must match for a True response.

Parameters:	client – An `elasticsearch.Elasticsearch` client object

curator.utils.is_master_node(client)¶

Return True if the connected client node is the elected master node in the Elasticsearch cluster, otherwise return False.

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	bool

curator.utils.name_to_node_id(client, name)¶

Return the node_id of the node identified by name

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	str

curator.utils.node_id_to_name(client, node_id)¶

Return the name of the node identified by node_id

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	str

curator.utils.node_roles(client, node_id)¶

Return the list of roles assigned to the node identified by node_id

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	list

curator.utils.parse_date_pattern(name)¶

Scan and parse name for time.strftime() strings, replacing them with the associated value when found, but otherwise returning lowercase values, as uppercase snapshot names are not allowed. It will detect if the first character is a <, which would indicate name is going to be using Elasticsearch date math syntax, and skip accordingly.

The time.strftime() identifiers that Curator currently recognizes as acceptable include:

Y: A 4 digit year
y: A 2 digit year
m: The 2 digit month
W: The 2 digit week of the year
d: The 2 digit day of the month
H: The 2 digit hour of the day, in 24 hour notation
M: The 2 digit minute of the hour
S: The 2 digit number of second of the minute
j: The 3 digit day of the year

Parameters:	name – A name, which can contain `time.strftime()` strings

curator.utils.parse_datemath(client, value)¶: Check if value is datemath. Parse it if it is. Return the bare value otherwise.

curator.utils.prune_nones(mydict)¶

Remove keys from mydict whose values are None

Parameters:	mydict – The dictionary to act on
Return type:	dict

curator.utils.read_file(myfile)¶

Read a file and return the resulting data.

Parameters:	myfile – A file to read.
Return type:	str

curator.utils.relocate_check(client, index)¶

This function calls client.cluster.state with a given index to check if all of the shards for that index are in the STARTED state. It will return True if all shards both primary and replica are in the STARTED state, and it will return False if any primary or replica shard is in a different state.

Parameters:	client – An `elasticsearch.Elasticsearch` client object index – The index to check the index shards state.

curator.utils.report_failure(exception)¶

Raise a exceptions.FailedExecution exception and include the original error message.

Parameters:	exception – The upstream exception.
Return type:	None

curator.utils.repository_exists(client, repository=None)¶

Verify the existence of a repository

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use
Return type:	bool

curator.utils.restore_check(client, index_list)¶

This function calls client.indices.recovery with the list of indices to check for complete recovery. It will return True if recovery of those indices is complete, and False otherwise. It is designed to fail fast: if a single shard is encountered that is still recovering (not in DONE stage), it will immediately return False, rather than complete iterating over the rest of the response.

Parameters:	client – An `elasticsearch.Elasticsearch` client object index_list – The list of indices to verify having been restored.

curator.utils.rollable_alias(client, alias)¶

Ensure that alias is an alias, and points to an index that can use the _rollover API.

Parameters:	client – An `elasticsearch.Elasticsearch` client object alias – An Elasticsearch alias

curator.utils.safe_to_snap(client, repository=None, retry_interval=120, retry_count=3)¶

Ensure there are no snapshots in progress. Pause and retry accordingly

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use retry_interval – Number of seconds to delay betwen retries. Default: 120 (seconds) retry_count – Number of attempts to make. Default: 3
Return type:	bool

curator.utils.show_dry_run(ilo, action, **kwargs)¶

Log dry run output with the action which would have been executed.

Parameters:	ilo – A `curator.indexlist.IndexList` action – The action to be performed. kwargs – Any other args to show in the log output

curator.utils.single_data_path(client, node_id)¶

In order for a shrink to work, it should be on a single filesystem, as shards cannot span filesystems. Return True if the node has a single filesystem, and False otherwise.

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	bool

curator.utils.snapshot_check(client, snapshot=None, repository=None)¶

This function calls client.snapshot.get and tests to see whether the snapshot is complete, and if so, with what status. It will log errors according to the result. If the snapshot is still IN_PROGRESS, it will return False. SUCCESS will be an INFO level message, PARTIAL nets a WARNING message, FAILED is an ERROR, message, and all others will be a WARNING level message.

Parameters:	client – An `elasticsearch.Elasticsearch` client object snapshot – The name of the snapshot. repository – The Elasticsearch snapshot repository to use

curator.utils.snapshot_in_progress(client, repository=None, snapshot=None)¶

Determine whether the provided snapshot in repository is IN_PROGRESS. If no value is provided for snapshot, then check all of them. Return snapshot if it is found to be in progress, or False

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use snapshot – The snapshot name

curator.utils.snapshot_running(client)¶

Return True if a snapshot is in progress, and False if not

Parameters:	client – An `elasticsearch.Elasticsearch` client object
Return type:	bool

curator.utils.task_check(client, task_id=None)¶

This function calls client.tasks.get with the provided task_id. If the task data contains 'completed': True, then it will return True If the task is not completed, it will log some information about the task and return False

Parameters:	client – An `elasticsearch.Elasticsearch` client object task_id – A task_id which ostensibly matches a task searchable in the tasks API.

curator.utils.test_client_options(config)¶

Test whether a SSL/TLS files exist. Will raise an exception if the files cannot be read.

Parameters:	config – A client configuration file data dictionary
Return type:	None

curator.utils.test_repo_fs(client, repository=None)¶

Test whether all nodes have write access to the repository

Parameters:	client – An `elasticsearch.Elasticsearch` client object repository – The Elasticsearch snapshot repository to use

curator.utils.to_csv(indices)¶

Return a csv string from a list of indices, or a single value if only one value is present

Parameters:	indices – A list of indices to act on, or a single value, which could be in the format of a csv string already.
Return type:	str

curator.utils.validate_actions(data)¶

Validate an Action configuration dictionary, as imported from actions.yml, for example.

The method returns a validated and sanitized configuration dictionary.

Parameters:	data – The configuration dictionary
Return type:	dict

curator.utils.validate_filters(action, filters)¶

Validate that the filters are appropriate for the action type, e.g. no index filters applied to a snapshot list.

Parameters:	action – An action name filters – A list of filters to test.

curator.utils.verify_client_object(test)¶

Test if test is a proper elasticsearch.Elasticsearch client object and raise an exception if it is not.

Parameters:	test – The variable or object to test
Return type:	None

curator.utils.verify_index_list(test)¶

Test if test is a proper curator.indexlist.IndexList object and raise an exception if it is not.

Parameters:	test – The variable or object to test
Return type:	None

curator.utils.verify_snapshot_list(test)¶

Test if test is a proper curator.snapshotlist.SnapshotList object and raise an exception if it is not.

Parameters:	test – The variable or object to test
Return type:	None

curator.utils.wait_for_it(client, action, task_id=None, snapshot=None, repository=None, index=None, index_list=None, wait_interval=9, max_wait=-1)¶

This function becomes one place to do all wait_for_completion type behaviors

Parameters:

client – An elasticsearch.Elasticsearch client object
action – The action name that will identify how to wait
task_id – If the action provided a task_id, this is where it must be declared.
snapshot – The name of the snapshot.
repository – The Elasticsearch snapshot repository to use
wait_interval – How frequently the specified “wait” behavior will be polled to check for completion.
max_wait – Number of seconds will the “wait” behavior persist before giving up and raising an Exception. The default is -1, meaning it will try forever.

class curator.SchemaCheck(config, schema, test_what, location)¶

Validate config with the provided voluptuous schema. test_what and location are for reporting the results, in case of failure. If validation is successful, the method returns config as valid.

Parameters:	config (dict) – A configuration dictionary. schema (`voluptuous.Schema`) – A voluptuous schema definition test_what (str) – which configuration block is being validated location (str) – An string to report which configuration sub-block is being tested.

Examples¶

Each of these examples presupposes that the requisite modules have been imported and an instance of the Elasticsearch client object has been created:

import elasticsearch
import curator

client = elasticsearch.Elasticsearch()

Filter indices by prefix¶

ilo = curator.IndexList(client)
ilo.filter_by_regex(kind='prefix', value='logstash-')

The contents of ilo.indices would then only be indices matching the prefix.

Filter indices by suffix¶

ilo = curator.IndexList(client)
ilo.filter_by_regex(kind='suffix', value='-prod')

The contents of ilo.indices would then only be indices matching the suffix.

Filter indices by age (name)¶

This example will match indices with the following criteria:

Have a date string of %Y.%m.%d
Use days as the unit of time measurement
Filter indices older than 5 days

ilo = curator.IndexList(client)
ilo.filter_by_age(source='name', direction='older', timestring='%Y.%m.%d',
    unit='days', unit_count=5
)

The contents of ilo.indices would then only be indices matching these criteria.

Filter indices by age (creation_date)¶

This example will match indices with the following criteria:

Use months as the unit of time measurement
Filter indices where the index creation date is older than 2 months from this moment.

ilo = curator.IndexList(client)
ilo.filter_by_age(source='creation_date', direction='older',
    unit='months', unit_count=2
)

The contents of ilo.indices would then only be indices matching these criteria.

Filter indices by age (field_stats)¶

This example will match indices with the following criteria:

Use days as the unit of time measurement
Filter indices where the timestamp field’s min_value is a date older than 3 weeks from this moment.

ilo = curator.IndexList(client)
ilo.filter_by_age(source='field_stats', direction='older',
    unit='weeks', unit_count=3, field='timestamp', stats_result='min_value'
)

The contents of ilo.indices would then only be indices matching these criteria.

Changelog¶

5.6.0 (13 November 2018)¶

New

The empty filter has been exposed for general use. This filter matches indices with no documents. (jrask) #1264

Added tests for Elasticsearch 6.3 and 6.4 releases. (untergeek)

Sort indices alphabetically before sorting by age. (tschroeder-zendesk) #1290

Add shards filtertype (cushind) #1298

Bug Fixes

Fix YAML linting so that YAML errors are caught and displayed on the command line. Reported in #1237 (untergeek)

Pin click version for compatibility. (Andrewsville) #1280

Allow much older epoch timestamps (rsteneteg) #1296

Reindex action respects ignore_empty_list flag (untergeek) #1297

Update ILM index version minimum to 6.6.0 (untergeek)

Catch reindex failures properly. Reported in #1260 (untergeek)

Documentation

Added Reindex example to the sidebar. (Nostalgiac) #1227

Fix Rollover example text and typos. (untergeek)

5.5.4 (23 May 2018)¶

Bug Fix

Extra args in show.py prevented show_snapshots from executing (untergeek)

5.5.3 (21 May 2018)¶

Short release cycle here specifically to address the Snapshot restore issue raised in #1192

Changes

By default, filter out indices with index.lifecycle.name set. This can be overridden with the option allow_ilm_indices with the caveat that you are on your own if there are conflicts. NOTE: The Index Lifecycle Management feature will not appear in Elasticsearch until 6.4.0

Removed some unused files from the repository.

Bug Fixes

Fix an ambiguously designed Alias test (untergeek)

Snapshot action will now raise an exception if the snapshot does not complete with state SUCCESS. Reported in #1192 (untergeek)

The show_indices and show_snapshots singletons were not working within the new framework. They’ve been fixed now.

5.5.2 (14 May 2018)¶

Changes

The alias, restore, rollover, and shrink actions have been added to curator_cli, along with a revamped method to manage/add actions in the future.

Updated certifi dependency to 2018.4.16

Added six dependency

Permit the use of versions 6.1 and greater of the elasticsearch python module. There are issues with SSL contexts in the 6.0 release that prevent Curator from being able to use this version. Currently the requirement version string is elasticsearch>=5.5.2,!=6.0.0,<7.0.0

Start of pylint cleanup, and use of six string_types. (untergeek)

Bug Fixes

unit_count_pattern setting can cause indices to mistakenly be included in an index filter. Fixed in #1206 (soenkeliebau)

Fix rollover _check_max_size() call. Reported in #1202 by @diranged (untergeek).

Update tested versions of Elasticsearch. (untergeek).

Update setup.cfg to install dependencies during source install. (untergeek)

Fix reference to unset variable name in log output at https://github.com/elastic/curator/blob/v5.5.1/curator/actions.py#L2145 It should be idx instead of index. (untergeek).

Alias action should raise NoIndices exception if warn_if_no_indices is True, and no add or remove sub-actions are found, rather than raising an ActionError. Reported in #1209 (untergeek).

Documentation

Clarify inclusive filtering for allocated filter. Fixed in #1203 (geekpete)

Fix Kibana filter description. #1199 (quartett-opa)

Add missing documentation about the new_name option for rollover. Reported in #1197 (untergeek)

5.5.1 (22 March 2018)¶

Bug Fixes

Fix pip installation issues for older versions of Python #1183 (untergeek)

5.5.0 (21 March 2018)¶

New Features

Add wait_for_rebalance as an option for shrink action. By default the behavior remains unchanged. You can now set this to False though to allow the shrink action to only check that the index being shrunk has finished being relocated and it will not wait for the cluster to rebalance. #1129 (tschroeder-zendesk)

Work around for extremely large cluster states. #1142 (rewiko)

Add CI tests for Elasticsearch versions 6.1 and 6.2 (untergeek)

Add Elasticsearch datemath support for snapshot names #1078 (untergeek)

Support max_size as a rollover condition for Elasticsearch versions 6.1.0 and up. #1140 (untergeek)

Skip indices with a document count of 0 when using source: field_stats to do age or period type filtering. #1130 (untergeek)

Bug Fixes

Fix missing node information in log line. #1142 (untergeek)

Fix default options in code that were causing schema validation errors after voluptuous upgrade to 0.11.1. Reported in #1149, fixed in #1156 (untergeek)

Disallow empty lists as reindex source. Raise exception if that happens. Reported in #1139 (untergeek)

Set a timeout_override for delete_snapshots to catch cases where slower repository network and/or disk access can cause a snapshot delete to take longer than the default 30 second client timeout. #1133 (untergeek)

Add AWS ES 5.1 support. #1172 (wanix)

Add missing period filter arguments for delete_snapshots. Reported in #1173 (untergeek)

Fix kibana filtertype to catch newer index names. Reported in #1171 (untergeek)

Re-order the closed indices filter for the Replicas action to take place before the empty list check. Reported in #1180 by @agomerz (untergeek)

General

Deprecate testing for Python 3.4. It is no longer being supported by Python.

Increase logging to show error when master_only is true and there are multiple hosts.

Documentation

Correct a misunderstanding about the nature of rollover conditions. #1144 (untergeek)

Correct links to the field_stats API, as it is non-existent in Elasticsearch 6.x. (untergeek)

Add a warning about using forcemerge on active indices. #1153 (untergeek)

Fix select URLs in pip installation from source to not be 404 #1133 (untergeek)

Fix an error in regex filter documentation #1138 (arne-cl)

5.4.1 (6 December 2017)¶

Bug Fixes

Improve Dockerfile to build from source and produce slimmer image #1111 (mikn)

Fix filter_kibana to correctly use exclude argument #1116 (cjuroz)

Fix ssl_no_validate behavior within AWS ES #1118 (igalarzab)

Improve command-line exception management #1119 (4383)

Make alias action always process remove before add to prevent undesired alias removals. #1120 (untergeek)

General

Bump ES versions in Travis CI

Documentation

Remove unit_count parameter doc for parameter that no longer exists #1107 (dashford)

Add missing exclude: True in timestring docs #1117 (GregMefford)

5.4.0 (13 November 2017)¶

Announcement

Support for Elasticsearch 6.0!!! Yes!

New Features

The field_stats API may be gone from Elasticsearch, but its utility cannot be denied. And so, Curator has replaced the field_stats API call with a small aggregation query. This will be perhaps a bit more costly in performance terms, as this small aggregation query must be made to each index in sequence, rather than as a one-shot call, like the field_stats API call. But the benefit will remain available, and it’s the only major API that did not persevere between Elasticsearch 5.x and 6.x that was needed by Curator.

5.3.0 (31 October 2017)¶

New Features

With the period filter and field_stats, it is useful to match indices that fit within the period, rather than just their start dates. This is now possible with intersect. See more in the documentation. Requested in #1045. (untergeek)

Add a restore function to curator_cli singleton. Mentioned in #851 (alexef)

Add pattern to the count filter. This is particularly useful when working with rollover indices. Requested in #1044 (untergeek)

The es_repo_mgr create command now can take skip_repo_fs_check as an argument (default is False) #1072 (alexef)

Add pattern_type feature expansion to the period filter. The default behavior is pattern_type='relative', which preserves existing behaviors so users with existing configurations can continue to use them without interruption. The new pattern_type is absolute, which allows you to specify hard dates for date_from and date_to, while date_from_format and date_to_format are strftime strings to interpret the from and to dates. Requested in #1047 (untergeek)

Add copy_aliases option to the shrink action. So this option is only set in the shrink action. The default value of the option is copy_aliases: 'False' and it does nothing. If you set to copy_aliases: 'True', you could copy the aliases from the source index to the target index. Requested in #1060 (monkey3199)

IAM Credentials can now be retrieved from the environment using the Boto3 Credentials provider. #1084 (kobuskc)

Bug Fixes

Delete the target index (if it exists) in the event that a shrink fails. Requested in #1058 (untergeek)

Fixed an integration test that could fail in the waning days of a month.

Fix build system anomalies for both unix and windows.

Documentation

Set repository access to be https by default.

Add documentation for copy_aliases option.

5.2.0 (1 September 2017)¶

New Features

Shrink action! Apologies to all who have patiently waited for this feature. It’s been a long time coming, but it is hopefully worth the wait. There are a lot of checks and tests associated with this action, as there are many conditions that have to be met in order for a shrink to take place. Curator will try its best to ensure that all of these conditions are met so you can comfortably rest assured that shrink will work properly unattended. See the documentation for more information.

The cli function has been split into cli and run functions. The behavior of cli will be indistinguishable from previous releases, preserving API integrity. The new run function allows lambda and other users to run Curator from the API with only a client configuration file and action file as arguments. Requested in #1031 (untergeek)

Allow use of time/date string interpolation for Rollover index naming. Added in #1010 (tschroeder-zendesk)

New unit_count_pattern allows you to derive the unit_count from the index name itself. This involves regular expressions, so be sure to do lots of testing in --dry-run mode before deploying to production. Added by (soenkeliebau) in #997

Bug Fixes

Reindex request_body allows for 2 different size options. One limits the number of documents reindexed. The other is for batch sizing. The batch sizing option was missing from the schema validator. This has been corrected. Reported in #1038 (untergeek)

A few sundry logging and notification changes were made.

5.1.2 (08 August 2017)¶

Errata

An update to Elasticsearch 5.5.0 changes the behavior of filter_by_aliases, differing from previous 5.x versions.

If a list of aliases is provided, indices must appear in all listed aliases or a 404 error will result, leading to no indices being matched. In older versions, if the index was associated with even one of the aliases in aliases, it would result in a match.

Tests and documentation have been updated to address these changes.

Debian 9 changed SSL versions, which means that the pre-built debian packages no longer work in Debian 9. In the short term, this requires a new repository. In the long term, I will try to get a better repository system working for these so they all work together, better. Requested in #998 (untergeek)

Bug Fixes

Support date math in reindex operations better. It did work previously, but would report failure because the test was looking for the index with that name from a list of indices, rather than letting Elasticsearch do the date math. Reported by DPattee in #1008 (untergeek)

Under rare circumstances, snapshot delete (or create) actions could fail, even when there were no snapshots in state IN_PROGRESS. This was tracked down by JD557 as a collision with a previously deleted snapshot that hadn’t finished deleting. It could be seen in the tasks API. An additional test for snapshot activity in the tasks API has been added to cover this scenario. Reported in #999 (untergeek)

The restore_check function did not work properly with wildcard index patterns. This has been rectified, and an integration test added to satisfy this. Reported in #989 (untergeek)

Make Curator report the Curator version, and not just reiterate the elasticsearch version when reporting version incompatibilities. Reported in #992. (untergeek)

Fix repository/snapshot name logging issue. #1005 (jpcarey)

Fix Windows build issue #1014 (untergeek)

Documentation

Fix/improve rST API documentation.

Thanks to many users who not only found and reported documentation issues, but also submitted corrections.

5.1.1 (8 June 2017)¶

Bug Fixes

Mock and cx_Freeze don’t play well together. Packages weren’t working, so I reverted the string-based comparison as before.

5.1.0 (8 June 2017)¶

New Features

Index Settings are here! First requested as far back as #160, it’s been requested in various forms culminating in #656. The official documentation addresses the usage. (untergeek)

Remote reindex now adds the ability to migrate from one cluster to another, preserving the index names, or optionally adding a prefix and/or a suffix. The official documentation shows you how. (untergeek)

Added support for naming rollover indices. #970 (jurajseffer)

Testing against ES 5.4.1, 5.3.3

Bug Fixes

Since Curator no longer supports old versions of python, convert tests to use isinstance. #973 (untergeek)

Fix stray instance of is not comparison instead of != #972 (untergeek)

Increase remote client timeout to 180 seconds for remote reindex. #930 (untergeek)

General

elasticsearch-py dependency bumped to 5.4.0

Added mock dependency due to isinstance and testing requirements

AWS ES 5.3 officially supports Curator now. Documentation has been updated to reflect this.

5.0.4 (16 May 2017)¶

Bug Fixes

The _recovery check needs to compare using != instead of is not, which apparently does not accurately compare unicode strings. Reported in #966 (untergeek)

5.0.3 (15 May 2017)¶

Bug Fixes

Restoring a snapshot on an exceptionally fast cluster/node can create a race condition where a _recovery check returns an empty dictionary {}, which causes Curator to fail. Added test and code to correct this. Reported in #962. (untergeek)

5.0.2 (4 May 2017)¶

Bug Fixes

Nasty bug in schema validation fixed where boolean options or filter flags would validate as True if non-boolean types were submitted. Reported in #945. (untergeek)

Check for presence of alias after reindex, in case the reindex was to an alias. Reported in #941. (untergeek)

Fix an edge case where an index named with 1970.01.01 could not be sorted by index-name age. Reported in #951. (untergeek)

Update tests to include ES 5.3.2

Bump certifi requirement to 2017.4.17.

Documentation

Document substitute strftime symbols for doing ISO Week timestrings added in #932. (untergeek)

Document how to include file paths better. Fixes #944. (untergeek)

5.0.1 (10 April 2017)¶

Bug Fixes

Fixed default values for include_global_state on the restore action to be in line with defaults in Elasticsearch 5.3

Documentation

Huge improvement to documenation, with many more examples.

Address age filter limitations per #859 (untergeek)

Address date matching behavior better per #858 (untergeek)

5.0.0 (5 April 2017)¶

The full feature set of 5.0 (including alpha releases) is included here.

New Features

Reindex is here! The new reindex action has a ton of flexibility. You can even reindex from remote locations, so long as the remote cluster is Elasticsearch 1.4 or newer.

Added the period filter (#733). This allows you to select indices or snapshots, based on whether they fit within a period of hours, days, weeks, months, or years.

Add dedicated “wait for completion” functionality. This supports health checks, recovery (restore) checks, snapshot checks, and operations which support the new tasks API. All actions which can use this have been refactored to take advantage of this. The benefit of this new feature is that client timeouts will be less likely to happen when performing long operations, like snapshot and restore.

NOTE: There is one caveat: forceMerge does not support this, per the Elasticsearch API. A forceMerge call will hold the client until complete, or the client times out. There is no clean way around this that I can discern.

Elasticsearch date math naming is supported and documented for the create_index action. An integration test is included for validation.

Allow allocation action to unset a key/value pair by using an empty value. Requested in #906. (untergeek)

Added support for the Rollover API. Requested in #898, and by countless others.

Added warn_if_no_indices option for alias action in response to #883. Using this option will permit the alias add or remove to continue with a logged warning, even if the filters result in a NoIndices condition. Use with care.

General

Bumped click (python module) version dependency to 6.7

Bumped urllib3 (python module) version dependency to 1.20

Bumped elasticsearch (python module) version dependency to 5.3

Refactored a ton of code to be cleaner and hopefully more consistent.

Bug Fixes

Curator now logs version incompatibilities as an error, rather than just raising an Exception. #874 (untergeek)

The get_repository() function now properly raises an exception instead of returning False if nothing is found. #761 (untergeek)

Check if an index is in an alias before attempting to delete it from the alias. Issue raised in #887. (untergeek)

Fix allocation issues when using Elasticsearch 5.1+. Issue raised in #871 (untergeek)

Documentation

Add missing repository arg to auto-gen API docs. Reported in #888 (untergeek)

Add all new documentation and clean up for v5 specific.

Breaking Changes

IndexList no longer checks to see if there are indices on initialization.

5.0.0a1 (23 March 2017)¶

This is the first alpha release of Curator 5. This should not be used for production! There will be many more changes before 5.0.0 is released.

New Features

Allow allocation action to unset a key/value pair by using an empty value. Requested in #906. (untergeek)

Added support for the Rollover API. Requested in #898, and by countless others.

Added warn_if_no_indices option for alias action in response to #883. Using this option will permit the alias add or remove to continue with a logged warning, even if the filters result in a NoIndices condition. Use with care.

Bug Fixes

Check if an index is in an alias before attempting to delete it from the alias. Issue raised in #887. (untergeek)

Fix allocation issues when using Elasticsearch 5.1+. Issue raised in #871 (untergeek)

Documentation

Add missing repository arg to auto-gen API docs. Reported in #888 (untergeek)

4.2.6 (27 January 2016)¶

General

Update Curator to use version 5.1 of the elasticsearch-py python module. With this change, there will be no reverse compatibility with Elasticsearch 2.x. For 2.x versions, continue to use the 4.x branches of Curator.

Tests were updated to reflect the changes in API calls, which were minimal.

Remove “official” support for Python 2.6. If you must use Curator on a system that uses Python 2.6 (RHEL/CentOS 6 users), it is recommended that you use the official RPM package as it is a frozen binary built on Python 3.5.x which will not conflict with your system Python.

Use isinstance() to verify client object. #862 (cp2587)

Prune older versions from Travis CI tests.

Update certifi dependency to latest version

Documentation

Add version compatibility section to official documentation.

Update docs to reflect changes. Remove cruft and references to older versions.

4.2.5 (22 December 2016)¶

General

Add and increment test versions for Travis CI. #839 (untergeek)

Make filter_list optional in snapshot, show_snapshot and show_indices singleton actions. #853 (alexef)

Bug Fixes

Fix cli integration test when different host/port are specified. Reported in #843 (untergeek)

Catch empty list condition during filter iteration in singleton actions. Reported in #848 (untergeek)

Documentation

Add docs regarding how filters are ANDed together, and how to do an OR with the regex pattern filter type. Requested in #842 (untergeek)

Fix typo in Click version in docs. #850 (breml)

Where applicable, replace [source,text] with [source,yaml] for better formatting in the resulting docs.

4.2.4 (7 December 2016)¶

Bug Fixes

--wait_for_completion should be True by default for Snapshot singleton action. Reported in #829 (untergeek)

Increase version_max to 5.1.99. Prematurely reported in #832 (untergeek)

Make the ‘.security’ index visible for snapshots so long as proper credentials are used. Reported in #826 (untergeek)

4.2.3.post1 (22 November 2016)¶

This fix is only going in for pip-based installs. There are no other code changes.

Bug Fixes

Fixed incorrect assumption of PyPI picking up dependency for certifi. It is still a dependency, but should not affect pip installs with an error any more. Reported in #821 (untergeek)

4.2.3 (21 November 2016)¶

4.2.2 was pulled immediately after release after it was discovered that the Windows binary distributions were still not including the certifi-provided certificates. This has now been remedied.

General

certifi is now officially a requirement.

setup.py now forcibly includes the certifi certificate PEM file in the “frozen” distributions (i.e., the compiled versions). The get_client method was updated to reflect this and catch it for both the Linux and Windows binary distributions. This should finally put to rest #810

4.2.2 (21 November 2016)¶

Bug Fixes

The certifi-provided certificates were not propagating to the compiled RPM/DEB packages. This has been corrected. Reported in #810 (untergeek)

General

Added missing --ignore_empty_list option to singleton actions. Requested in #812 (untergeek)

Documentation

Add a FAQ entry regarding the click module’s need for Unicode when using Python 3. Kind of a bug fix too, as the entry_points were altered to catch this omission and report a potential solution on the command-line. Reported in #814 (untergeek)

Change the “Command-Line” documentation header to be “Running Curator”

4.2.1 (8 November 2016)¶

Bug Fixes

In the course of package release testing, an undesirable scenario was caught where boolean flags default values for curator_cli were improperly overriding values from a yaml config file.

General

Adding in direct download URLs for the RPM, DEB, tarball and zip packages.

4.2.0 (4 November 2016)¶

New Features

Shard routing allocation enable/disable. This will allow you to disable shard allocation routing before performing one or more actions, and then re-enable after it is complete. Requested in #446 (untergeek)

Curator 3.x-style command-line. This is now curator_cli, to differentiate between the current binary. Not all actions are available, but the most commonly used ones are. With the addition in 4.1.0 of schema and configuration validation, there’s even a way to still do filter chaining on the command-line! Requested in #767, and by many other users (untergeek)

General

Update testing to the most recent versions.

Lock elasticsearch-py module version at >= 2.4.0 and <= 3.0.0. There are API changes in the 5.0 release that cause tests to fail.

Bug Fixes

Guarantee that binary packages are built from the latest Python + libraries. This ensures that SSL/TLS will work without warning messages about insecure connections, unless they actually are insecure. Reported in #780, though the reported problem isn’t what was fixed. The fix is needed based on what was discovered while troubleshooting the problem. (untergeek)

4.1.2 (6 October 2016)¶

This release does not actually add any new code to Curator, but instead improves documentation and includes new linux binary packages.

General

New Curator binary packages for common Linux systems! These will be found in the same repositories that the python-based packages are in, but have no dependencies. All necessary libraries/modules are bundled with the binary, so everything should work out of the box. This feature doesn’t change any other behavior, so it’s not a major release.

These binaries have been tested in:

CentOS 6 & 7

Ubuntu 12.04, 14.04, 16.04

Debian 8

They do not work in Debian 7 (library mismatch). They may work in other systems, but that is untested.

The script used is in the unix_packages directory. The Vagrantfiles for the various build systems are in the Vagrant directory.

Bug Fixes

The only bug that can be called a bug is actually a stray .exe suffix in the binary package creation section (cx_freeze) of setup.py. The Windows binaries should have .exe extensions, but not unix variants.

Elasticsearch 5.0.0-beta1 testing revealed that a document ID is required during document creation in tests. This has been fixed, and a redundant bit of code in the forcemerge integration test was removed.

Documentation

The documentation has been updated and improved. Examples and installation are now top-level events, with the sub-sections each having their own link. They also now show how to install and use the binary packages, and the section on installation from source has been improved. The missing section on installing the voluptuous schema verification module has been written and included. #776 (untergeek)

4.1.1 (27 September 2016)¶

Bug Fixes

String-based booleans are now properly coerced. This fixes an issue where True/False were used in environment variables, but not recognized. #765 (untergeek)

Fix missing count method in __map_method in SnapshotList. Reported in #766 (untergeek)

General

Update es_repo_mgr to use the same client/logging YAML config file. Requested in #752 (untergeek)

Schema Validation

Cases where source was not defined in a filter (but should have been) were informing users that a timestring field was there that shouldn’t have been. This edge case has been corrected.

Documentation

Added notifications and FAQ entry to explain that AWS ES is not supported.

4.1.0 (6 September 2016)¶

New Features

Configuration and Action file schema validation. Requested in #674 (untergeek)

Alias filtertype! With this filter, you can select indices based on whether they are part of an alias. Merged in #748 (untergeek)

Count filtertype! With this filter, you can now configure Curator to only keep the most recent n indices (or snapshots!). Merged in #749 (untergeek)

Experimental! Use environment variables in your YAML configuration files. This was a popular request, #697. (untergeek)

General

New requirement! voluptuous Python schema validation module

Requirement version bump: Now requires elasticsearch-py 2.4.0

Bug Fixes

delete_aliases option in close action no longer results in an error if not all selected indices have an alias. Add test to confirm expected behavior. Reported in #736 (untergeek)

Documentation

Add information to FAQ regarding indices created before Elasticsearch 1.4. Merged in #747

4.0.6 (15 August 2016)¶

Bug Fixes

Update old calls used with ES 1.x to reflect changes in 2.x+. This was necessary to work with Elasticsearch 5.0.0-alpha5. Fixed in #728 (untergeek)

Doc Fixes

Add section detailing that the value of a value filter element should be encapsulated in single quotes. Reported in #726. (untergeek)

4.0.5 (3 August 2016)¶

Bug Fixes

Fix incorrect variable name for AWS Region reported in #679 (basex)

Fix filter_by_space() to not fail when index age metadata is not present. Indices without the appropriate age metadata will instead be excluded, with a debug-level message. Reported in #724 (untergeek)

Doc Fixes

Fix documentation for the space filter and the source filter element.

4.0.4 (1 August 2016)¶

Bug Fixes

Fix incorrect variable name in Allocation action. #706 (lukewaite)

Incorrect error message in create_snapshot_body reported in #711 (untergeek)

Test for empty index list object should happen in action initialization for snapshot action. Discovered in #711. (untergeek)

Doc Fixes

Add menus to asciidoc chapters #704 (untergeek)

Add pyyaml dependency #710 (dtrv)

4.0.3 (22 July 2016)¶

General

4.0.2 didn’t work for pip installs due to an omission in the MANIFEST.in file. This came up during release testing, but before the release was fully published. As the release was never fully published, this should not have actually affected anyone.

Bug Fixes

These are the same as 4.0.2, but it was never fully released.

All default settings are now values returned from functions instead of constants. This was resulting in settings getting stomped on. New test addresses the original complaint. This removes the need for deepcopy. See issue #687 (untergeek)

Fix host vs. hosts issue in get_client() rather than the non-functional function in repomgrcli.py.

Update versions being tested.

Community contributed doc fixes.

Reduced logging verbosity by making most messages debug level. #684 (untergeek)

Fixed log whitelist behavior (and switched to blacklisting instead). Default behavior will now filter traffic from the elasticsearch and urllib3 modules.

Fix Travis CI testing to accept some skipped tests, as needed. #695 (untergeek)

Fix missing empty index test in snapshot action. #682 (sherzberg)

4.0.2 (22 July 2016)¶

Bug Fixes

All default settings are now values returned from functions instead of constants. This was resulting in settings getting stomped on. New test addresses the original complaint. This removes the need for deepcopy. See issue #687 (untergeek)

Fix host vs. hosts issue in get_client() rather than the non-functional function in repomgrcli.py.

Update versions being tested.

Community contributed doc fixes.

Reduced logging verbosity by making most messages debug level. #684 (untergeek)

Fixed log whitelist behavior (and switched to blacklisting instead). Default behavior will now filter traffic from the elasticsearch and urllib3 modules.

Fix Travis CI testing to accept some skipped tests, as needed. #695 (untergeek)

Fix missing empty index test in snapshot action. #682 (sherzberg)

4.0.1 (1 July 2016)¶

Bug Fixes

Coerce Logstash/JSON logformat type timestamp value to always use UTC. #661 (untergeek)

Catch and remove indices from the actionable list if they do not have a creation_date field in settings. This field was introduced in ES v1.4, so that indicates a rather old index. #663 (untergeek)

Replace missing state filter for snapshotlist. #665 (untergeek)

Restore es_repo_mgr as a stopgap until other CLI scripts are added. It will remain undocumented for now, as I am debating whether to make repository creation its own action in the API. #668 (untergeek)

Fix dry run results for snapshot action. #673 (untergeek)

4.0.0 (24 June 2016)¶

It’s official! Curator 4.0.0 is released!

Breaking Changes

New and improved API!

Command-line changes. No more command-line args, except for --config, --actions, and --dry-run:

--config points to a YAML client and logging configuration file. The default location is ~/.curator/curator.yml

--actions arg points to a YAML action configuration file

--dry-run will simulate the action(s) which would have taken place, but not actually make any changes to the cluster or its indices.

New Features

Snapshot restore is here!

YAML configuration files. Now a single file can define an entire batch of commands, each with their own filters, to be performed in sequence.

Sort by index age not only by index name (as with previous versions of Curator), but also by index creation_date, or by calculations from the Field Stats API on a timestamp field.

Atomically add/remove indices from aliases! This is possible by way of the new IndexList class and YAML configuration files.

State of indices pulled and stored in IndexList instance. Fewer API calls required to serially test for open/close, size_in_bytes, etc.

Filter by space now allows sorting by age!

Experimental! Use AWS IAM credentials to sign requests to Elasticsearch. This requires the end user to manually install the requests_aws4auth python module.

Optionally delete aliases from indices before closing.

An empty index or snapshot list no longer results in an error if you set ignore_empty_list to True. If True it will still log that the action was not performed, but will continue to the next action. If ‘False’ it will log an ERROR and exit with code 1.

API

Updated API documentation

Class: IndexList. This pulls all indices at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

Class: SnapshotList. This pulls all snapshots from the given repository at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

Add wait_for_completion to Allocation and Replicas actions. These will use the client timeout, as set by default or timeout_override, to determine how long to wait for timeout. These are handled in batches of indices for now.

Allow timeout_override option for all actions. This allows for different timeout values per action.

Improve API by giving each action its own do_dry_run() method.

General

Updated use documentation for Elastic main site.

Include example files for --config and --actions.

4.0.0b2 (16 June 2016)¶

Second beta release of the 4.0 branch

New Feature

An empty index or snapshot list no longer results in an error if you set ignore_empty_list to True. If True it will still log that the action was not performed, but will continue to the next action. If ‘False’ it will log an ERROR and exit with code 1. (untergeek)

4.0.0b1 (13 June 2016)¶

First beta release of the 4.0 branch!

The release notes will be rehashing the new features in 4.0, rather than the bug fixes done during the alphas.

Breaking Changes

New and improved API!

Command-line changes. No more command-line args, except for --config, --actions, and --dry-run:

--config points to a YAML client and logging configuration file. The default location is ~/.curator/curator.yml

--actions arg points to a YAML action configuration file

--dry-run will simulate the action(s) which would have taken place, but not actually make any changes to the cluster or its indices.

New Features

Snapshot restore is here!

YAML configuration files. Now a single file can define an entire batch of commands, each with their own filters, to be performed in sequence.

Sort by index age not only by index name (as with previous versions of Curator), but also by index creation_date, or by calculations from the Field Stats API on a timestamp field.

Atomically add/remove indices from aliases! This is possible by way of the new IndexList class and YAML configuration files.

State of indices pulled and stored in IndexList instance. Fewer API calls required to serially test for open/close, size_in_bytes, etc.

Filter by space now allows sorting by age!

Experimental! Use AWS IAM credentials to sign requests to Elasticsearch. This requires the end user to manually install the requests_aws4auth python module.

Optionally delete aliases from indices before closing.

API

Updated API documentation

Class: IndexList. This pulls all indices at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

Class: SnapshotList. This pulls all snapshots from the given repository at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

Add wait_for_completion to Allocation and Replicas actions. These will use the client timeout, as set by default or timeout_override, to determine how long to wait for timeout. These are handled in batches of indices for now.

Allow timeout_override option for all actions. This allows for different timeout values per action.

Improve API by giving each action its own do_dry_run() method.

General

Updated use documentation for Elastic main site.

Include example files for --config and --actions.

4.0.0a10 (10 June 2016)¶

New Features

Snapshot restore is here!

Optionally delete aliases from indices before closing. Fixes #644 (untergeek)

General

Add wait_for_completion to Allocation and Replicas actions. These will use the client timeout, as set by default or timeout_override, to determine how long to wait for timeout. These are handled in batches of indices for now.

Allow timeout_override option for all actions. This allows for different timeout values per action.

Bug Fixes

Disallow use of master_only if multiple hosts are used. Fixes #615 (untergeek)

Fix an issue where arguments weren’t being properly passed and populated.

ForceMerge replaced Optimize in ES 2.1.0.

Fix prune_nones to work with Python 2.6. Fixes #619 (untergeek)

Fix TimestringSearch to work with Python 2.6. Fixes #622 (untergeek)

Add language classifiers to setup.py. Fixes #640 (untergeek)

Changed references to readthedocs.org to be readthedocs.io.

4.0.0a9 (27 Apr 2016)¶

General

Changed create_index API to use kwarg extra_settings instead of body

Normalized Alias action to use name instead of alias. This simplifies documentation by reducing the number of option elements.

Streamlined some code

Made exclude a filter element setting for all filters. Updated all examples to show this.

Improved documentation

New Features

Alias action can now accept extra_settings to allow adding filters, and/or routing.

4.0.0a8 (26 Apr 2016)¶

Bug Fixes

Fix to use optimize with versions of Elasticsearch < 5.0

Fix missing setting in testvars

4.0.0a7 (25 Apr 2016)¶

Bug Fixes

Fix AWS4Auth error.

4.0.0a6 (25 Apr 2016)¶

General

Documentation updates.

Improve API by giving each action its own do_dry_run() method.

Bug Fixes

Do not escape characters other than . and - in timestrings. Fixes #602 (untergeek)

** New Features**

Added CreateIndex action.

4.0.0a4 (21 Apr 2016)¶

Bug Fixes

Require pyyaml 3.10 or better.

In the case that no options are in an action, apply the defaults.

4.0.0a3 (21 Apr 2016)¶

It’s time for Curator 4.0 alpha!

Breaking Changes

New API! (again?!)

Command-line changes. No more command-line args, except for --config, --actions, and --dry-run:

--config points to a YAML client and logging configuration file. The default location is ~/.curator/curator.yml

--actions arg points to a YAML action configuration file

--dry-run will simulate the action(s) which would have taken place, but not actually make any changes to the cluster or its indices.

General

Updated API documentation

Updated use documentation for Elastic main site.

Include example files for --config and --actions.

New Features

Sort by index age not only by index name (as with previous versions of Curator), but also by index creation_date, or by calculations from the Field Stats API on a timestamp field.

Class: IndexList. This pulls all indices at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

Class: SnapshotList. This pulls all snapshots from the given repository at instantiation, and you apply filters, which are class methods. You can iterate over as many filters as you like, in fact, due to the YAML config file.

YAML configuration files. Now a single file can define an entire batch of commands, each with their own filters, to be performed in sequence.

Atomically add/remove indices from aliases! This is possible by way of the new IndexList class and YAML configuration files.

State of indices pulled and stored in IndexList instance. Fewer API calls required to serially test for open/close, size_in_bytes, etc.

Filter by space now allows sorting by age!

Experimental! Use AWS IAM credentials to sign requests to Elasticsearch. This requires the end user to manually install the requests_aws4auth python module.

3.5.1 (21 March 2016)¶

Bug fixes

Add more logging information to snapshot delete method #582 (untergeek)

Improve default timeout, logging, and exception handling for seal command #583 (untergeek)

Fix use of default snapshot name. #584 (untergeek)

3.5.0 (16 March 2016)¶

General

Add support for the –client-cert and –client-key command line parameters and client_cert and client_key parameters to the get_client() call. #520 (richm)

Bug fixes

Disallow users from creating snapshots with upper-case letters, which is not permitted by Elasticsearch. #562 (untergeek)

Remove print() command from setup.py as it causes issues with command-line retrieval of --url, etc. #568 (thib-ack)

Remove unnecessary argument from build_filter() #530 (zzugg)

Allow day of year filter to be made up with 1, 2 or 3 digits #578 (petitout)

3.4.1 (10 February 2016)¶

General

Update license copyright to 2016

Use slim python version with Docker #527 (xaka)

Changed --master-only exit code to 0 when connected to non-master node #540 (wkruse)

Add cx_Freeze capability to setup.py, plus a binary_release.py script to simplify binary package creation. #554 (untergeek)

Set Elastic as author. #555 (untergeek)

Put repository creation methods into API and document them. Requested in #550 (untergeek)

Bug fixes

Fix sphinx documentation build error #506 (hydrapolic)

Ensure snapshots are found before iterating #507 (garyelephant)

Fix a doc inconsistency #509 (pmoust)

Fix a typo in show documentation #513 (pbamba)

Default to trying the cluster state for checking whether indices are closed, and then fall back to using the _cat API (for Amazon ES instances). #519 (untergeek)

Improve logging to show time delay between optimize runs, if selected. #525 (untergeek)

Allow elasticsearch-py module versions through 2.3.0 (a presumption at this point) #524 (untergeek)

Improve logging in snapshot api method to reveal when a repository appears to be missing. Reported in #551 (untergeek)

Test that --timestring has the correct variable for --time-unit. Reported in #544 (untergeek)

Allocation will exit with exit_code 0 now when there are no indices to work on. Reported in #531 (untergeek)

3.4.0 (28 October 2015)¶

General

API change in elasticsearch-py 1.7.0 prevented alias operations. Fixed in #486 (HonzaKral)

During index selection you can now select only closed indices with --closed-only. Does not impact --all-indices Reported in #476. Fixed in #487 (Basster)

API Changes in Elasticsearch 2.0.0 required some refactoring. All tests pass for ES versions 1.0.3 through 2.0.0-rc1. Fixed in #488 (untergeek)

es_repo_mgr now has access to the same SSL options from #462. #489 (untergeek)

Logging improvements requested in #475. (untergeek)

Added --quiet flag. #494 (untergeek)

Fixed index_closed to work with AWS Elasticsearch. #499 (univerio)

Acceptable versions of Elasticsearch-py module are 1.8.0 up to 2.1.0 (untergeek)

3.3.0 (31 August 2015)¶

Announcement

Curator is tested in Jenkins. Each commit to the master branch is tested with both Python versions 2.7.6 and 3.4.0 against each of the following Elasticsearch versions: * 1.7_nightly * 1.6_nightly * 1.7.0 * 1.6.1 * 1.5.1 * 1.4.4 * 1.3.9 * 1.2.4 * 1.1.2 * 1.0.3

If you are using a version different from this, your results may vary.

General

Allocation type can now also be include or exclude, in addition to the existing default require type. Add --type to the allocation command to specify the type. #443 (steffo)

Bump elasticsearch python module dependency to 1.6.0+ to enable synced_flush API call. Reported in #447 (untergeek)

Add SSL features, --ssl-no-validate and certificate to provide other ways to validate SSL connections to Elasticsearch. #436 (untergeek)

Bug fixes

Delete by space was only reporting space used by primary shards. Fixed to show all space consumed. Reported in #455 (untergeek)

Update exit codes and messages for snapshot selection. Reported in #452 (untergeek)

Fix potential int/float casting issues. Reported in #465 (untergeek)

3.2.3 (16 July 2015)¶

Bug fix

In order to address customer and community issues with bulk deletes, the master_timeout is now invoked for delete operations. This should address 503s with 30s timeouts in the debug log, even when --timeout is set to a much higher value. The master_timeout is tied to the --timeout flag value, but will not exceed 300 seconds. #420 (untergeek)

General

Mixing it up a bit here by putting General second! The only other changes are that logging has been improved for deletes so you won’t need to have the --debug flag to see if you have error codes >= 400, and some code documentation improvements.

3.2.2 (13 July 2015)¶

General

This is a very minor change. The mock library recently removed support for Python 2.6. As many Curator users are using RHEL/CentOS 6, which is pinned to Python 2.6, this requires the mock version referenced by Curator to also be pinned to a supported version (mock==1.0.1).

3.2.1 (10 July 2015)¶

General

Added delete verification & retry (fixed at 3x) to potentially cover an edge case in #420 (untergeek)

Since GitHub allows rST (reStructuredText) README documents, and that’s what PyPI wants also, the README has been rebuilt in rST. (untergeek)

Bug fixes

If closing indices with ES 1.6+, and all indices are closed, ensure that the seal command does not try to seal all indices. Reported in #426 (untergeek)

Capture AttributeError when sealing indices if a non-TransportError occurs. Reported in #429 (untergeek)

3.2.0 (25 June 2015)¶

New!

Added support to manually seal, or perform a [synced flush](http://www.elastic.co/guide/en/elasticsearch/reference/current/indices-synced-flush.html) on indices with the seal command. #394 (untergeek)

Added experimental support for SSL certificate validation. In order for this to work, you must install the certifi python module: pip install certifi This feature should automatically work if the certifi module is installed. Please report any issues.

General

Changed logging to go to stdout rather than stderr. Reopened #121 and figured they were right. This is better. (untergeek)

Exit code 99 was unpopular. It has been removed. Reported in #371 and #391 (untergeek)

Add --skip-repo-validation flag for snapshots. Do not validate write access to repository on all cluster nodes before proceeding. Useful for shared filesystems where intermittent timeouts can affect validation, but won’t likely affect snapshot success. Requested in #396 (untergeek)

An alias no longer needs to be pre-existent in order to use the alias command. #317 (untergeek)

es_repo_mgr now passes through upstream errors in the event a repository fails to be created. Requested in #405 (untergeek)

Bug fixes

In rare cases, * wildcard would not expand. Replaced with _all. Reported in #399 (untergeek)

Beginning with Elasticsearch 1.6, closed indices cannot have their replica count altered. Attempting to do so results in this error: org.elasticsearch.ElasticsearchIllegalArgumentException: Can't update [index.number_of_replicas] on closed indices [[test_index]] - can leave index in an unopenable state As a result, the change_replicas method has been updated to prune closed indices. This change will apply to all versions of Elasticsearch. Reported in #400 (untergeek)

Fixed es_repo_mgr repository creation verification error. Reported in #389 (untergeek)

3.1.0 (21 May 2015)¶

General

If wait_for_completion is true, snapshot success is now tested and logged. Reported in #253 (untergeek)

Log & return false if a snapshot is already in progress (untergeek)

Logs individual deletes per index, even though they happen in batch mode. Also log individual snapshot deletions. Reported in #372 (untergeek)

Moved chunk_index_list from cli to api utils as it’s now also used by filter.py

Added a warning and 10 second timer countdown if you use --timestring to filter indices, but do not use --older-than or --newer-than in conjunction with it. This is to address #348, which behavior isn’t a bug, but prevents accidental action against all of your time-series indices. The warning and timer are not displayed for show and --dry-run operations.

Added tests for es_repo_mgr in #350

Doc fixes

Bug fixes

delete-by-space needed the same fix used for #245. Fixed in #353 (untergeek)

Increase default client timeout for es_repo_mgr as node discovery and availability checks for S3 repositories can take a bit. Fixed in #352 (untergeek)

If an index is closed, indicate in show and --dry-run output. Reported in #327. (untergeek)

Fix issue where CLI parameters were not being passed to the es_repo_mgr create sub-command. Reported in #337. (feltnerm)

3.0.3 (27 Mar 2015)¶

Announcement

This is a bug fix release. #319 and #320 are affecting a few users, so this release is being expedited.

Test count: 228 Code coverage: 99%

General

Documentation for the CLI converted to Asciidoc and moved to http://www.elastic.co/guide/en/elasticsearch/client/curator/current/index.html

Improved logging, and refactored a few methods to help with this.

Dry-run output is now more like v2, with the index or snapshot in the log line, along with the command. Several tests needed refactoring with this change, along with a bit of documentation.

Bug fixes

Fix links to repository in setup.py. Reported in #318 (untergeek)

No more --delay with optimized indices. Reported in #319 (untergeek)

--request_timeout not working as expected. Reinstate the version 2 timeout override feature to prevent default timeouts for optimize and snapshot operations. Reported in #320 (untergeek)

Reduce index count to 200 for test.integration.test_cli_commands.TestCLISnapshot.test_cli_snapshot_huge_list in order to reduce or eliminate Jenkins CI test timeouts. Reported in #324 (untergeek)

--dry-run no longer calls show, but will show output in the log, as in v2. This was a recurring complaint. See #328 (untergeek)

3.0.2 (23 Mar 2015)¶

Announcement

This is a bug fix release. #307 and #309 were big enough to warrant an expedited release.

Bug fixes

Purge unneeded constants, and clean up config options for snapshot. Reported in #303 (untergeek)

Don’t split large index list if performing snapshots. Reported in #307 (untergeek)

Act correctly if a zero value for –older-than or –newer-than is provided. #309 (untergeek)

3.0.1 (16 Mar 2015)¶

Announcement

The regex_iterate method was horribly named. It has been renamed to apply_filter. Methods have been added to allow API users to build a filtered list of indices similarly to how the CLI does. This was an oversight. Props to @SegFaultAX for pointing this out.

General

In conjunction with the rebrand to Elastic, URLs and documentation were updated.

Renamed horribly named regex_iterate method to apply_filter #298 (untergeek)

Added build_filter method to mimic CLI calls. #298 (untergeek)

Added Examples page in the API documentation. #298 (untergeek)

Bug fixes

Refactored to show –dry-run info for –disk-space calls. Reported in #290 (untergeek)

Added list chunking so acting on huge lists of indices won’t result in a URL bigger than 4096 bytes (Elasticsearch’s default limit.) Reported in https://github.com/elastic/curator/issues/245#issuecomment-77916081

Refactored to_csv() method to be simpler.

Added and removed tests according to changes. Code coverage still at 99%

3.0.0 (9 March 2015)¶

Release Notes

The full release of Curator 3.0 is out! Check out all of the changes here!

Note: This release is _not_ reverse compatible with any previous version.

Because 3.0 is a major point release, there have been some major changes to both the API as well as the CLI arguments and structure.

Be sure to read the updated command-line specific docs in the [wiki](https://github.com/elasticsearch/curator/wiki) and change your command-line arguments accordingly.

The API docs are still at http://curator.readthedocs.io. Be sure to read the latest docs, or select the docs for 3.0.0.

General

Breaking changes to the API. Because this is a major point revision, changes to the API have been made which are non-reverse compatible. Before upgrading, be sure to update your scripts and test them thoroughly.

Python 3 support Somewhere along the line, Curator would no longer work with curator. All tests now pass for both Python2 and Python3, with 99% code coverage in both environments.

New CLI library. Using Click now. http://click.pocoo.org/3/ This change is especially important as it allows very easy CLI integration testing.

Pipelined filtering! You can now use --older-than & --newer-than in the same command! You can also provide your own regex via the --regex parameter. You can use multiple instances of the --exclude flag.

Manually include indices! With the --index paramter, you can add an index to the working list. You can provide multiple instances of the --index parameter as well!

Tests! So many tests now. Test coverage of the API methods is at 100% now, and at 99% for the CLI methods. This doesn’t mean that all of the tests are perfect, or that I haven’t missed some scenarios. It does mean, however, that it will be much easier to write tests if something turns up missed. It also means that any new functionality will now need to have tests.

Iteration changes Methods now only iterate through each index when appropriate! In fact, the only commands that iterate are alias and optimize. The bloom command will iterate, but only if you have added the –delay flag with a value greater than zero.

Improved packaging! Methods have been moved into categories of api and cli, and further broken out into individual modules to help them be easier to find and read.

Check for allocation before potentially re-applying an allocation rule. #273 (ferki)

Assigning replica count and routing allocation rules _can_ be done to closed indices. #283 (ferki)

Bug fixes

Don’t accidentally delete .kibana index. #261 (malagoli)

Fix segment count for empty indices. #265 (untergeek)

Change bloom filter cutoff Elasticsearch version to 1.4. Reported in #267 (untergeek)

3.0.0rc1 (5 March 2015)¶

Release Notes

RC1 is here! I’m re-releasing the Changes from all betas here, minus the intra-beta code fixes. Barring any show stoppers, the official release will be soon.

General

Breaking changes to the API. Because this is a major point revision, changes to the API have been made which are non-reverse compatible. Before upgrading, be sure to update your scripts and test them thoroughly.

Python 3 support Somewhere along the line, Curator would no longer work with curator. All tests now pass for both Python2 and Python3, with 99% code coverage in both environments.

New CLI library. Using Click now. http://click.pocoo.org/3/ This change is especially important as it allows very easy CLI integration testing.

Pipelined filtering! You can now use --older-than & --newer-than in the same command! You can also provide your own regex via the --regex parameter. You can use multiple instances of the --exclude flag.

Manually include indices! With the --index paramter, you can add an index to the working list. You can provide multiple instances of the --index parameter as well!

Tests! So many tests now. Test coverage of the API methods is at 100% now, and at 99% for the CLI methods. This doesn’t mean that all of the tests are perfect, or that I haven’t missed some scenarios. It does mean, however, that it will be much easier to write tests if something turns up missed. It also means that any new functionality will now need to have tests.

Methods now only iterate through each index when appropriate!

Improved packaging! Hopefully the entry_point issues some users have had will be addressed by this. Methods have been moved into categories of api and cli, and further broken out into individual modules to help them be easier to find and read.

Check for allocation before potentially re-applying an allocation rule. #273 (ferki)

Assigning replica count and routing allocation rules _can_ be done to closed indices. #283 (ferki)

Bug fixes

Don’t accidentally delete .kibana index. #261 (malagoli)

Fix segment count for empty indices. #265 (untergeek)

Change bloom filter cutoff Elasticsearch version to 1.4. Reported in #267 (untergeek)

3.0.0b4 (5 March 2015)¶

Notes

Integration testing! Because I finally figured out how to use the Click Testing API, I now have a good collection of command-line simulations, complete with a real back-end. This testing found a few bugs (this is why testing exists, right?), and fixed a few of them.

Bug fixes

HUGE! curator show snapshots would _delete_ snapshots. This is fixed.

Return values are now being sent from the commands.

scripttest is no longer necessary (click.Test works!)

Calling get_snapshot without a snapshot name returns all snapshots

3.0.0b3 (4 March 2015)¶

Bug fixes

setup.py was lacking the new packages “curator.api” and “curator.cli” The package works now.

Python3 suggested I had to normalize the beta tag to just b3, so that’s also changed.

Cleaned out superfluous imports and logger references from the __init__.py files.

3.0.0-beta2 (3 March 2015)¶

Bug fixes

Python3 issues resolved. Tests now pass on both Python2 and Python3

3.0.0-beta1 (3 March 2015)¶

General

Breaking changes to the API. Because this is a major point revision, changes to the API have been made which are non-reverse compatible. Before upgrading, be sure to update your scripts and test them thoroughly.

New CLI library. Using Click now. http://click.pocoo.org/3/

Pipelined filtering! You can now use --older-than & --newer-than in the same command! You can also provide your own regex via the --regex parameter. You can use multiple instances of the --exclude flag.

Manually include indices! With the --index paramter, you can add an index to the working list. You can provide multiple instances of the --index parameter as well!

Tests! So many tests now. Unit test coverage of the API methods is at 100% now. This doesn’t mean that all of the tests are perfect, or that I haven’t missed some scenarios. It does mean that any new functionality will need to also have tests, now.

Methods now only iterate through each index when appropriate!

Improved packaging! Hopefully the entry_point issues some users have had will be addressed by this. Methods have been moved into categories of api and cli, and further broken out into individual modules to help them be easier to find and read.

Check for allocation before potentially re-applying an allocation rule. #273 (ferki)

Bug fixes

Don’t accidentally delete .kibana index. #261 (malagoli)

Fix segment count for empty indices. #265 (untergeek)

Change bloom filter cutoff Elasticsearch version to 1.4. Reported in #267 (untergeek)

2.1.2 (22 January 2015)¶

Bug fixes

Do not try to set replica count if count matches provided argument. #247 (bobrik)

Fix JSON logging (Logstash format). #250 (magnusbaeck)

Fix bug in filter_by_space() which would match all indices if the provided patterns found no matches. Reported in #254 (untergeek)

2.1.1 (30 December 2014)¶

Bug fixes

Renamed unnecessarily redundant --replicas to --count in args for curator_script.py

2.1.0 (30 December 2014)¶

General

Snapshot name now appears in log output or STDOUT. #178 (untergeek)

Replicas! You can now change the replica count of indices. Requested in #175 (untergeek)

Delay option added to Bloom Filter functionality. #206 (untergeek)

Add 2-digit years as acceptable pattern (y vs. Y). Reported in #209 (untergeek)

Add Docker container definition #226 (christianvozar)

Allow the use of 0 with –older-than, –most-recent and –delete-older-than. See #208. #211 (bobrik)

Bug fixes

Edge case where 1.4.0.Beta1-SNAPSHOT would break version check. Reported in #183 (untergeek)

Typo fixed. #193 (ferki)

Type fixed. #204 (gheppner)

Shows proper error in the event of concurrent snapshots. #177 (untergeek)

Fixes erroneous index display of _, a, l, l when –all-indices selected. Reported in #222 (untergeek)

Use json.dumps() to escape exceptions. Reported in #210 (untergeek)

Check if index is closed before adding to alias. Reported in #214 (bt5e)

No longer force-install argparse if pre-installed #216 (whyscream)

Bloom filters have been removed from Elasticsearch 1.5.0. Update methods and tests to act accordingly. #233 (untergeek)

2.0.2 (8 October 2014)¶

Bug fixes

Snapshot name not displayed in log or STDOUT #185 (untergeek)

Variable name collision in delete_snapshot() #186 (untergeek)

2.0.1 (1 October 2014)¶

Bug fix

Override default timeout when snapshotting –all-indices #179 (untergeek)

2.0.0 (25 September 2014)¶

General

New! Separation of Elasticsearch Curator Python API and curator_script.py (untergeek)

New! --delay after optimize to allow cluster to quiesce #131 (untergeek)

New! --suffix option in addition to --prefix #136 (untergeek)

New! Support for wildcards in prefix & suffix #136 (untergeek)

Complete refactor of snapshots. Now supporting incrementals! (untergeek)

Bug fix

Incorrect error msg if no indices sent to create_snapshot (untergeek)

Correct for API change coming in ES 1.4 #168 (untergeek)

Missing " in Logstash log format #143 (cassianoleal)

Change non-master node test to exit code 0, log as INFO. #145 (untergeek)

months option missing from validate_timestring() (untergeek)

1.2.2 (29 July 2014)¶

Bug fix

Updated README.md to briefly explain what curator does #117 (untergeek)

Fixed es_repo_mgr logging whitelist #119 (untergeek)

Fixed absent months time-unit #120 (untergeek)

Filter out .marvel-kibana when prefix is .marvel- #120 (untergeek)

Clean up arg parsing code where redundancy exists #123 (untergeek)

Properly divide debug from non-debug logging #125 (untergeek)

Fixed show command bug caused by changes to command structure #126 (michaelweiser)

1.2.1 (24 July 2014)¶

Bug fix

Fixed the new logging when called by curator entrypoint.

1.2.0 (24 July 2014)¶

General

New! Allow user-specified date patterns: --timestring #111 (untergeek)

New! Curate weekly indices (must use week of year) #111 (untergeek)

New! Log output in logstash format --logformat logstash #111 (untergeek)

Updated! Cleaner default logs (debug still shows everything) (untergeek)

Improved! Dry runs are more visible in log output (untergeek)

Errata

The --separator option was removed in lieu of user-specified date patterns.

Default --timestring for days: %Y.%m.%d (Same as before)

Default --timestring for hours: %Y.%m.%d.%H (Same as before)

Default --timestring for weeks: %Y.%W

1.1.3 (18 July 2014)¶

Bug fix

Prefix not passed in get_object_list() #106 (untergeek)

Use os.devnull instead of /dev/null for Windows #102 (untergeek)

The http auth feature was erroneously omitted #100 (bbuchacher)

1.1.2 (13 June 2014)¶

Bug fix

This was a showstopper bug for anyone using RHEL/CentOS with a Python 2.6 dependency for yum

Python 2.6 does not like format calls without an index. #96 via #95 (untergeek)

We won’t talk about what happened to 1.1.1. No really. I hate git today :(

1.1.0 (12 June 2014)¶

General

Updated! New command structure

New! Snapshot to fs or s3 #82 (untergeek)

New! Add/Remove indices to alias #82 via #86 (cschellenger)

New! --exclude-pattern #80 (ekamil)

New! (sort of) Restored --log-level support #73 (xavier-calland)

New! show command-line options #82 via #68 (untergeek)

New! Shard Allocation Routing #82 via #62 (nickethier)

Bug fix

Fix --max_num_segments not being passed correctly #74 (untergeek)

Change BUILD_NUMBER to CURATOR_BUILD_NUMBER in setup.py #60 (mohabusama)

Fix off-by-one error in time calculations #66 (untergeek)

Fix testing with python3 #92 (untergeek)

Errata

Removed optparse compatibility. Now requires argparse.

1.0.0 (25 Mar 2014)¶

General

compatible with elasticsearch-py 1.0 and Elasticsearch 1.0 (honzakral)

Lots of tests! (honzakral)

Streamline code for 1.0 ES versions (honzakral)

Bug fix

Fix find_expired_indices() to not skip closed indices (honzakral)

0.6.2 (18 Feb 2014)¶

General

Documentation fixes #38 (dharrigan)

Add support for HTTPS URI scheme and optparse compatibility for Python 2.6 (gelim)

Add elasticsearch module version checking for future compatibility checks (untergeek)

0.6.1 (08 Feb 2014)¶

General

Added tarball versioning to setup.py (untergeek)

Bug fix

Fix long_description by including README.md in MANIFEST.in (untergeek)

Incorrect version number in curator.py (untergeek)

0.6.0 (08 Feb 2014)¶

General

Restructured repository to a be a proper python package. (arieb)

Added setup.py file. (arieb)

Removed the deprecated file logstash_index_cleaner.py (arieb)

Updated README.md to fit the new package, most importantly the usage and installation. (arieb)

Fixes and package push to PyPI (untergeek)

0.5.2 (26 Jan 2014)¶

General

Fix boolean logic determining hours or days for time selection (untergeek)

0.5.1 (20 Jan 2014)¶

General

Fix can_bloom to compare numbers (HonzaKral)

Switched find_expired_indices() to use datetime and timedelta

Do not try and catch unrecoverable exceptions. (HonzaKral)

Future proofing the use of the elasticsearch client (i.e. work with version 1.0+ of Elasticsearch) (HonzaKral) Needs more testing, but should work.

Add tests for these scenarios (HonzaKral)

0.5.0 (17 Jan 2014)¶

General

Deprecated logstash_index_cleaner.py Use new curator.py instead (untergeek)

new script change: curator.py (untergeek)

new add index optimization (Lucene forceMerge) to reduce segments and therefore memory usage. (untergeek)

update refactor of args and several functions to streamline operation and make it more readable (untergeek)

update refactor further to clean up and allow immediate (and future) portability (HonzaKral)

0.4.0¶

General

First version logged in CHANGELOG

new --disable-bloom-days feature requires 0.90.9+

http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/index-modules-codec.html#bloom-postings

This can save a lot of heap space on cold indexes (i.e. not actively indexing documents)

License¶

Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.