OpenSearch/docs/reference/index-modules/merge.asciidoc

[[index-modules-merge]]
== Merge

A shard in elasticsearch is a Lucene index, and a Lucene index is broken
down into segments. Segments are internal storage elements in the index
where the index data is stored, and are immutable up to delete markers.
Segments are, periodically, merged into larger segments to keep the
index size at bay and expunge deletes.

The more segments one has in the Lucene index means slower searches and
more memory used. Segment merging is used to reduce the number of segments,
however merges can be expensive to perform, especially on low IO environments.
Merges can be throttled using <<store-throttling,store level throttling>>.


[float]
[[policy]]
=== Policy

The index merge policy module allows one to control which segments of a
shard index are to be merged. There are several types of policies with
the default set to `tiered`.

[float]
[[tiered]]
==== tiered

Merges segments of approximately equal size, subject to an allowed
number of segments per tier. This is similar to `log_bytes_size` merge
policy, except this merge policy is able to merge non-adjacent segment,
and separates how many segments are merged at once from how many
segments are allowed per tier. This merge policy also does not
over-merge (i.e., cascade merges).

This policy has the following settings:

`index.merge.policy.expunge_deletes_allowed`::

	When expungeDeletes is called, we only merge away a segment if its delete
	percentage is over this threshold. Default is `10`.

`index.merge.policy.floor_segment`::

	Segments smaller than this are "rounded up" to this size, i.e. treated as
	equal (floor) size for merge selection. This is to prevent frequent
	flushing of tiny segments, thus preventing a long tail in the index. Default
	is `2mb`.

`index.merge.policy.max_merge_at_once`::

	Maximum number of segments to be merged at a time during "normal" merging.
	Default is `10`.

`index.merge.policy.max_merge_at_once_explicit`::

	Maximum number of segments to be merged at a time, during optimize or
	expungeDeletes. Default is `30`.

`index.merge.policy.max_merged_segment`::

	Maximum sized segment to produce during normal merging (not explicit
	optimize). This setting is approximate: the estimate of the merged segment
	size is made by summing sizes of to-be-merged segments (compensating for
	percent deleted docs). Default is `5gb`.

`index.merge.policy.segments_per_tier`::

	Sets the allowed number of segments per tier. Smaller values mean more
	merging but fewer segments. Default is `10`. Note, this value needs to be
	>= than the `max_merge_at_once` otherwise you'll force too many merges to
	occur.

`index.reclaim_deletes_weight`::

	Controls how aggressively merges that reclaim more deletions are favored.
	Higher values favor selecting merges that reclaim deletions. A value of
	`0.0` means deletions don't impact merge selection. Defaults to `2.0`.

`index.compound_format`::

	Should the index be stored in compound format or not. Defaults to `false`.
	See <<index-compound-format,`index.compound_format`>> in
	<<index-modules-settings>>.

For normal merging, this policy first computes a "budget" of how many
segments are allowed to be in the index. If the index is over-budget,
then the policy sorts segments by decreasing size (proportionally considering percent
deletes), and then finds the least-cost merge. Merge cost is measured by
a combination of the "skew" of the merge (size of largest seg divided by
smallest seg), total merge size and pct deletes reclaimed, so that
merges with lower skew, smaller size and those reclaiming more deletes,
are favored.

If a merge will produce a segment that's larger than
`max_merged_segment` then the policy will merge fewer segments (down to
1 at once, if that one has deletions) to keep the segment size under
budget.

Note, this can mean that for large shards that holds many gigabytes of
data, the default of `max_merged_segment` (`5gb`) can cause for many
segments to be in an index, and causing searches to be slower. Use the
indices segments API to see the segments that an index has, and
possibly either increase the `max_merged_segment` or issue an optimize
call for the index (try and aim to issue it on a low traffic time).

[float]
[[log-byte-size]]
==== log_byte_size

A merge policy that merges segments into levels of exponentially
increasing *byte size*, where each level has fewer segments than the
value of the merge factor. Whenever extra segments (beyond the merge
factor upper bound) are encountered, all segments within the level are
merged.

This policy has the following settings:

[cols="<,<",options="header",]
|=======================================================================
|Setting |Description
|index.merge.policy.merge_factor |Determines how often segment indices
are merged by index operation. With smaller values, less RAM is used
while indexing, and searches on unoptimized indices are faster, but
indexing speed is slower. With larger values, more RAM is used during
indexing, and while searches on unoptimized indices are slower, indexing
is faster. Thus larger values (greater than 10) are best for batch index
creation, and smaller values (lower than 10) for indices that are
interactively maintained. Defaults to `10`.

|index.merge.policy.min_merge_size |A size setting type which sets the
minimum size for the lowest level segments. Any segments below this size
are considered to be on the same level (even if they vary drastically in
size) and will be merged whenever there are mergeFactor of them. This
effectively truncates the "long tail" of small segments that would
otherwise be created into a single level. If you set this too large, it
could greatly increase the merging cost during indexing (if you flush
many small segments). Defaults to `1.6mb`

|index.merge.policy.max_merge_size |A size setting type which sets the
largest segment (measured by total byte size of the segment's files)
that may be merged with other segments. Defaults to unbounded.

|index.merge.policy.max_merge_docs |Determines the largest segment
(measured by document count) that may be merged with other segments.
Defaults to unbounded.
|=======================================================================

[float]
[[log-doc]]
==== log_doc

A merge policy that tries to merge segments into levels of exponentially
increasing *document count*, where each level has fewer segments than
the value of the merge factor. Whenever extra segments (beyond the merge
factor upper bound) are encountered, all segments within the level are
merged.

[cols="<,<",options="header",]
|=======================================================================
|Setting |Description
|index.merge.policy.merge_factor |Determines how often segment indices
are merged by index operation. With smaller values, less RAM is used
while indexing, and searches on unoptimized indices are faster, but
indexing speed is slower. With larger values, more RAM is used during
indexing, and while searches on unoptimized indices are slower, indexing
is faster. Thus larger values (greater than 10) are best for batch index
creation, and smaller values (lower than 10) for indices that are
interactively maintained. Defaults to `10`.

|index.merge.policy.min_merge_docs |Sets the minimum size for the lowest
level segments. Any segments below this size are considered to be on the
same level (even if they vary drastically in size) and will be merged
whenever there are mergeFactor of them. This effectively truncates the
"long tail" of small segments that would otherwise be created into a
single level. If you set this too large, it could greatly increase the
merging cost during indexing (if you flush many small segments).
Defaults to `1000`.

|index.merge.policy.max_merge_docs |Determines the largest segment
(measured by document count) that may be merged with other segments.
Defaults to unbounded.
|=======================================================================

[float]
[[scheduling]]
=== Scheduling

The merge scheduler (ConcurrentMergeScheduler) controls the execution of
merge operations once they are needed (according to the merge policy).  Merges
run in separate threads, and when the maximum number of threads is reached,
further merges will wait until a merge thread becomes available.  The merge
scheduler supports this setting:

`index.merge.scheduler.max_thread_count`::

The maximum number of threads that may be merging at once. Defaults to
`Math.max(1, Math.min(3, Runtime.getRuntime().availableProcessors() / 2))`
which works well for a good solid-state-disk (SSD).  If your index is on
spinning platter drives instead, decrease this to 1.

[float]
==== SerialMergeScheduler

A merge scheduler that simply does each merge sequentially using the
calling thread (blocking the operations that triggered the merge or the
index operation). This merge scheduler has a merge thread pool that
explicitly schedules merges, and it makes sure that merges are serial
within a shard, yet concurrent across multiple shards.

The scheduler supports the following settings:

`index.merge.scheduler.max_merge_at_once`::

The maximum number of merges a single merge run performs. This setting prevents
executing unlimited amount of merges in a loop until another shards has a
chance to get a merge thread from the pool. If this limit is reached the
merge thread returns to the pool and continues once the the call to a single
shards is executed. The default is `5`