Shrink Indexedit

The shrink index API allows you to shrink an existing index into a new index with fewer primary shards. The requested number of primary shards in the target index must be a factor of the number of shards in the source index. For example an index with 8 primary shards can be shrunk into 4, 2 or 1 primary shards or an index with 15 primary shards can be shrunk into 5, 3 or 1. If the number of shards in the index is a prime number it can only be shrunk into a single primary shard. Before shrinking, a (primary or replica) copy of every shard in the index must be present on the same node.

Shrinking works as follows:

  • First, it creates a new target index with the same definition as the source index, but with a smaller number of primary shards.
  • Then it hard-links segments from the source index into the target index. (If the file system doesn’t support hard-linking, then all segments are copied into the new index, which is a much more time consuming process.)
  • Finally, it recovers the target index as though it were a closed index which had just been re-opened.

Preparing an index for shrinkingedit

In order to shrink an index, the index must be marked as read-only, and a (primary or replica) copy of every shard in the index must be relocated to the same node and have health green.

These two conditions can be achieved with the following request:

PUT /my_source_index/_settings
{
  "settings": {
    "index.routing.allocation.require._name": "shrink_node_name", 
    "index.blocks.write": true 
  }
}

Forces the relocation of a copy of each shard to the node with name shrink_node_name. See Shard Allocation Filtering for more options.

Prevents write operations to this index while still allowing metadata changes like deleting the index.

It can take a while to relocate the source index. Progress can be tracked with the _cat recovery API, or the cluster health API can be used to wait until all shards have relocated with the wait_for_no_relocating_shards parameter.

Shrinking an indexedit

To shrink my_source_index into a new index called my_target_index, issue the following request:

POST my_source_index/_shrink/my_target_index?copy_settings=true
{
  "settings": {
    "index.routing.allocation.require._name": null, 
    "index.blocks.write": null 
  }
}

Clear the allocation requirement copied from the source index.

Clear the index write block copied from the source index.

The above request returns immediately once the target index has been added to the cluster state — it doesn’t wait for the shrink operation to start.

Indices can only be shrunk if they satisfy the following requirements:

  • the target index must not exist
  • The index must have more primary shards than the target index.
  • The number of primary shards in the target index must be a factor of the number of primary shards in the source index. The source index must have more primary shards than the target index.
  • The index must not contain more than 2,147,483,519 documents in total across all shards that will be shrunk into a single shard on the target index as this is the maximum number of docs that can fit into a single shard.
  • The node handling the shrink process must have sufficient free disk space to accommodate a second copy of the existing index.

The _shrink API is similar to the create index API and accepts settings and aliases parameters for the target index:

POST my_source_index/_shrink/my_target_index?copy_settings=true
{
  "settings": {
    "index.number_of_replicas": 1,
    "index.number_of_shards": 1, 
    "index.codec": "best_compression" 
  },
  "aliases": {
    "my_search_indices": {}
  }
}

The number of shards in the target index. This must be a factor of the number of shards in the source index.

Best compression will only take affect when new writes are made to the index, such as when force-merging the shard to a single segment.

Mappings may not be specified in the _shrink request.

By default, with the exception of index.analysis, index.similarity, and index.sort settings, index settings on the source index are not copied during a shrink operation. With the exception of non-copyable settings, settings from the source index can be copied to the target index by adding the URL parameter copy_settings=true to the request. Note that copy_settings can not be set to false. The parameter copy_settings will be removed in 8.0.0

[6.4.0] Deprecated in 6.4.0. not copying settings is deprecated, copying settings will be the default behavior in 7.x

Monitoring the shrink processedit

The shrink process can be monitored with the _cat recovery API, or the cluster health API can be used to wait until all primary shards have been allocated by setting the wait_for_status parameter to yellow.

The _shrink API returns as soon as the target index has been added to the cluster state, before any shards have been allocated. At this point, all shards are in the state unassigned. If, for any reason, the target index can’t be allocated on the shrink node, its primary shard will remain unassigned until it can be allocated on that node.

Once the primary shard is allocated, it moves to state initializing, and the shrink process begins. When the shrink operation completes, the shard will become active. At that point, Elasticsearch will try to allocate any replicas and may decide to relocate the primary shard to another node.

Wait For Active Shardsedit

Because the shrink operation creates a new index to shrink the shards to, the wait for active shards setting on index creation applies to the shrink index action as well.