Helpers

Date Ops

class curator.helpers.date_ops.TimestringSearch(timestring)

Bases: object

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

Parameters:

timestring (strftime()) – An strftime pattern

get_epoch(searchme)

Parameters:

searchme – A string to be matched against pattern that matches timestring

Returns:

The epoch timestamp extracted from searchme by regex matching against pattern

Return type:

int or None

pattern

Object attribute. re.compile(regex) where regex = r'(?P<date>{0})'.format(get_date_regex(timestring)). Uses get_date_regex()

timestring

Object attribute preserving param timestring

curator.helpers.date_ops.absolute_date_range(unit, date_from, date_to, date_from_format=None, date_to_format=None)

This function calculates a date range with an absolute time stamp for both the start time and the end time. These dates are converted to epoch time. The parameter unit is used when the same simplified date is used for both date_from and date_to to calculate the duration. 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.

Parameters:

• unit (str) – One of hours, days, weeks, months, or years.

• date_from (str) – The simplified date for the start of the range

• date_to (str) – The simplified date for the end of the range.

• date_from_format (str) – The strftime() string used to parse date_from

• date_to_format (str) – The strftime() string used to parse date_to

Returns:

The epoch start time and end time of a date range

Return type:

tuple

curator.helpers.date_ops.date_range(unit, range_from, range_to, epoch=None, week_starts_on='sunday')

This function calculates a date range with a distinct epoch time stamp of both the start time and the end time in counts of unit relative to the time at execution, if epoch is not set.

If unit is weeks, you can also determine when a week begins using week_starts_on, which can be either sunday or monday.

Parameters:

• unit (str) – One of hours, days, weeks, months, or years.

• range_from (int) – Count of unit in the past/future is the origin?

• range_to (int) – Count of unit in the past/future is the end point?

• epoch (int) – An epoch timestamp used to establish a point of reference for calculations.

• week_starts_on (str) – Either sunday or monday. Default is sunday

Returns:

The epoch start time and end time of a date range

Return type:

tuple

curator.helpers.date_ops.datetime_to_epoch(mydate)

Converts datetime into epoch seconds

Parameters:

mydate (datetime) – A Python datetime

Returns:

An epoch timestamp based on mydate

Return type:

int

curator.helpers.date_ops.fix_epoch(epoch)

Fix value of epoch to be the count since the epoch in seconds only, which should be 10 or fewer digits long.

Parameters:

epoch (int) – An epoch timestamp, in epoch + milliseconds, or microsecond, or even nanoseconds.

Returns:

An epoch timestamp in seconds only, based on epoch

Return type:

int

curator.helpers.date_ops.get_date_regex(timestring)

Parameters:

timestring (strftime()) – An strftime pattern

Returns:

A regex string based on a provided strftime() timestring.

Return type:

str

curator.helpers.date_ops.get_datemath(client, datemath, random_element=None)

Parameters:

• client (Elasticsearch) – A client connection object

• datemath (datemath) – An elasticsearch datemath string

• random_element (str) – This allows external randomization of the name and is only useful for tests so that you can guarantee the output because you provided the random portion.

Returns:

the parsed index name from datemath

Return type:

str

curator.helpers.date_ops.get_datetime(index_timestamp, timestring)

Parameters:

• index_timestamp (str) – The index timestamp

• timestring (strftime()) – An strftime pattern

Returns:

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

Return type:

datetime

curator.helpers.date_ops.get_point_of_reference(unit, count, epoch=None)

Parameters:

• unit (str) – One of seconds, minutes, hours, days, weeks, months, or years.

• unit_count (int) – The number of units. unit_count * unit will be calculated out to the relative number of seconds.

• epoch (int) – An epoch timestamp used in conjunction with unit and unit_count to establish a point of reference for calculations.

Returns:

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

Return type:

int

curator.helpers.date_ops.get_unit_count_from_name(index_name, pattern)

Parameters:

• index_name (str) – An index name

• pattern (str) – A regular expression pattern

Returns:

The unit count, if a match is able to be found in the name

Return type:

int

curator.helpers.date_ops.handle_iso_week_number(mydate, timestring, index_timestamp)

Parameters:

• mydate (datetime) – A Python datetime

• timestring (strftime()) – An strftime pattern

• index_timestamp (str) – The index timestamp

Returns:

The date of the previous week based on ISO week number

Return type:

datetime

curator.helpers.date_ops.isdatemath(data)

Parameters:

data (str) – An expression to validate as being datemath or not

Returns:

True if data is a valid datemath expression, else False

Return type:

bool

curator.helpers.date_ops.parse_date_pattern(name)

Scan and parse name for 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 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 (str) – A name, which can contain strftime() strings

Returns:

The parsed date pattern

Return type:

str

curator.helpers.date_ops.parse_datemath(client, value)

Validate that value looks like proper datemath. If it passes this test, then try to ship it to Elasticsearch for real. It may yet fail this test, and if it does, it will raise a ConfigurationError exception. If it passes, return the fully parsed string.

Parameters:

• client (Elasticsearch) – A client connection object

• value (str) – A string to check for datemath

Returns:

A datemath indexname, fully rendered by Elasticsearch

Return type:

str

Getters

curator.helpers.getters.byte_size(num, suffix='B')

Parameters:

• num (int) – The number of byte

• suffix (str) – An arbitrary suffix, like Bytes

Returns:

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

Return type:

float

curator.helpers.getters.get_alias_actions(oldidx, newidx, aliases)

Parameters:

• oldidx (str) – The old index name

• newidx (str) – The new index name

• aliases (dict) – The aliases

Returns:

A list of actions suitable for update_aliases() actions kwarg.

Return type:

list

curator.helpers.getters.get_data_tiers(client)

Get all valid data tiers from the node roles of each node in the cluster by polling each node

Parameters:

client (Elasticsearch) – A client connection object

Returns:

The available data tiers in tier: bool form.

Return type:

dict

curator.helpers.getters.get_indices(client, search_pattern='*', include_hidden=False)

Calls indices()

Will use the provided search_pattern to get a list of indices from the cluster. If include_hidden is True, it will include hidden indices in ‘expand_wildcards’.

:param

client: A client connection object search_pattern: The index search pattern to use include_hidden: Include hidden indices in the list

:type

client: Elasticsearch search_pattern: str include_hidden: bool

Returns:

The matching list of indices from the cluster

Return type:

list

curator.helpers.getters.get_repository(client, repository='')

Calls get_repository()

Parameters:

• client (Elasticsearch) – A client connection object

• repository (str) – The Elasticsearch snapshot repository to use

Returns:

Configuration information for repository.

Return type:

dict

curator.helpers.getters.get_snapshot(client, repository=None, snapshot='')

Calls get()

Parameters:

• client (Elasticsearch) – A client connection object

• repository (str) – The Elasticsearch snapshot repository to use

• snapshot (str) – The snapshot name, or a comma-separated list of snapshots

Returns:

Information about the provided snapshot, a snapshot (or a comma-separated list of snapshots). If no snapshot specified, it will collect info for all snapshots. If none exist, an empty dict will be returned.

Return type:

dict

curator.helpers.getters.get_snapshot_data(client, repository=None)

Get all snapshots from repository and return a list. Calls get()

Parameters:

• client (Elasticsearch) – A client connection object

• repository (str) – The Elasticsearch snapshot repository to use

Returns:

The list of all snapshots from repository

Return type:

list

curator.helpers.getters.get_tier_preference(client, target_tier='data_frozen')

Do the tier preference thing in reverse order from coldest to hottest Based on the value of target_tier, build out the list to use.

Parameters:

• client (Elasticsearch) – A client connection object

• target_tier (str) – The target data tier, e.g. data_warm.

Returns:

A suitable tier preference string in csv format

Return type:

str

curator.helpers.getters.get_write_index(client, alias)

Calls get_alias()

Parameters:

• client (Elasticsearch) – A client connection object

• alias (str) – An alias name

Returns:

The the index name associated with the alias that is designated is_write_index

Return type:

str

curator.helpers.getters.index_size(client, idx, value='total')

Calls stats()

Parameters:

• client (Elasticsearch) – A client connection object

• idx (str) – An index name

• value (str) – One of either primaries or total

Returns:

The sum of either primaries or total shards for index idx

Return type:

integer

curator.helpers.getters.name_to_node_id(client, name)

Calls info()

Parameters:

• client (Elasticsearch) – A client connection object

• name (str) – The node name

Returns:

The node_id of the node identified by name

Return type:

str

curator.helpers.getters.node_id_to_name(client, node_id)

Calls info()

Parameters:

• client (Elasticsearch) – A client connection object

• node_id (str) – The node node_id

Returns:

The name of the node identified by node_id

Return type:

str

curator.helpers.getters.node_roles(client, node_id)

Calls info()

Parameters:

• client (Elasticsearch) – A client connection object

• node_id (str) – The node node_id

Returns:

The list of roles assigned to the node identified by node_id

Return type:

list

curator.helpers.getters.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. Calls stats()

Parameters:

• client (Elasticsearch) – A client connection object

• node_id (str) – The node node_id

Returns:

True if the node has a single filesystem, else False

Return type:

bool

Testers

curator.helpers.testers.ilm_policy_check(client, alias)

Test if alias is associated with an ILM policy

Calls get_settings()

Parameters:

• client (Elasticsearch) – A client connection object

• alias (str) – The alias name

Return type:

bool

curator.helpers.testers.repository_exists(client, repository=None)

Calls get_repository()

Parameters:

• client (Elasticsearch) – A client connection object

• repository (str) – The Elasticsearch snapshot repository to use

Returns:

True if repository exists, else False

Return type:

bool

curator.helpers.testers.rollable_alias(client, alias)

Calls get_alias()

Parameters:

• client (Elasticsearch) – A client connection object

• alias (str) – An Elasticsearch alias

Returns:

True or False depending on whether alias is an alias that points to an index that can be used by the _rollover API.

Return type:

bool

curator.helpers.testers.snapshot_running(client)

Calls get_repository()

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

Parameters:

client (Elasticsearch) – A client connection object

Return type:

bool

curator.helpers.testers.validate_actions(data)

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

Parameters:

data (dict) – The configuration dictionary

Returns:

The validated and sanitized configuration dictionary.

Return type:

dict

curator.helpers.testers.validate_filters(action, myfilters)

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

Parameters:

• action (str) – An action name

• myfilters (list) – A list of filters to test.

Returns:

Validated list of filters

Return type:

list

curator.helpers.testers.verify_client_object(test)

Parameters:

test (Elasticsearch) – The variable or object to test

Returns:

True if test is a proper Elasticsearch client object and raise a TypeError exception if it is not.

Return type:

bool

curator.helpers.testers.verify_repository(client, repository=None)

Do verify_repository() call. If it fails, raise a RepositoryException.

Parameters:

• client (Elasticsearch) – A client connection object

• repository (str) – A repository name

Return type:

None

curator.helpers.testers.verify_snapshot_list(test)

Parameters:

test (SnapshotList) – The variable or object to test

Returns:

None if test is a proper SnapshotList object, else raise a TypeError exception.

Return type:

None

Utils

curator.helpers.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 (list) – The list of indices

Returns:

A list of lists (each a piece of the original indices)

Return type:

list

curator.helpers.utils.report_failure(exception)

Raise a FailedExecution exception and include the original error message.

Parameters:

exception (:py:exc:Exception) – The upstream exception.

Return type:

None

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

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

Parameters:

• ilo (IndexList) – An IndexList Object

• action (str) – The action to be performed.

• kwargs (dict) – Any other args to show in the log output

Return type:

None

curator.helpers.utils.to_csv(indices)

Parameters:

indices (list) – A list of indices to act on, or a single value, which could be in the format of a csv string already.

Returns:

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

Return type:

str

curator.helpers.utils.multitarget_fix(pattern: str) → str

If pattern only has ‘-’ prefixed entries (excludes) prepend a wildcard to pattern

Parameters:

pattern (str) – The Elasticsearch multi-target syntax pattern

Returns:

The pattern, possibly with a wildcard prepended

Return type:

str

curator.helpers.utils.regex_loop(matchstr: str, indices: list) → list

Loop through indices, Match against matchstr, return matches

Parameters:

• matchstr (str) – The Python regex pattern to match against

• indices (list) – The list of indices to match against

Returns:

The list of matching indices

Return type:

list

curator.helpers.utils.multitarget_match(pattern: str, index_list: list) → list

Convert Elasticsearch multi-target syntax pattern into Python regex patterns. Match against index_list and return the list of matches while excluding any negative matches.

Parameters:

• pattern (str) – The Elasticsearch multi-target syntax pattern

• index_list (list) – The list of indices to match against

Returns:

The final resulting list of indices

Return type:

list

Waiters

curator.helpers.waiters.health_check(client, **kwargs)

This function calls client.cluster. health() and, based on the params 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 (Elasticsearch) – A client connection object

Return type:

bool

curator.helpers.waiters.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 primary and replica shards are in the STARTED state, and it will return False if any shard is in a different state.

Parameters:

• client (Elasticsearch) – A client connection object

• index (str) – The index name

Return type:

bool

curator.helpers.waiters.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 (Elasticsearch) – A client connection object

• index_list (list) – The list of indices to verify having been restored.

• kwargs – Any additional keyword arguments to pass to the function

Return type:

bool

curator.helpers.waiters.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 (Elasticsearch) – A client connection object

• snapshot (str) – The snapshot name

• repository (str) – The repository name

Return type:

bool

curator.helpers.waiters.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 (Elasticsearch) – A client connection object

• task_id (str) – The task id

Return type:

bool

curator.helpers.waiters.wait_for_it(client, action, task_id=None, snapshot=None, repository=None, index=None, index_list=None, wait_interval=9, max_wait=-1, **kwargs)

This function becomes one place to do all wait_for_completion type behaviors

Parameters:

• client (Elasticsearch) – A client connection object

• action (str) – The action name that will identify how to wait

• task_id (str) – If the action provided a task_id, this is where it must be declared.

• snapshot (str) – The name of the snapshot.

• repository (str) – The Elasticsearch snapshot repository to use

• wait_interval (int) – Seconds to wait between completion checks.

• max_wait (int) – Maximum number of seconds to wait_for_completion

• kwargs (dict) – Any additional keyword arguments to pass to the function

Return type:

None