OpenSearch/docs/reference/cluster/reroute.asciidoc

[[cluster-reroute]]
== Cluster Reroute

The reroute command allows for manual changes to the allocation of individual
shards in the cluster. For example, a shard can be moved from one node to
another explicitly, an allocation can be cancelled, and an unassigned shard can
be explicitly allocated to a specific node.

Here is a short example of a simple reroute API call:

[source,js]
--------------------------------------------------
POST /_cluster/reroute
{
    "commands" : [
        {
            "move" : {
                "index" : "test", "shard" : 0,
                "from_node" : "node1", "to_node" : "node2"
            }
        },
        {
          "allocate_replica" : {
                "index" : "test", "shard" : 1,
                "node" : "node3"
          }
        }
    ]
}
--------------------------------------------------
// CONSOLE
// TEST[skip:doc tests run with only a single node]

It is important to note that that after processing any reroute commands
Elasticsearch will perform rebalancing as normal (respecting the values of
settings such as `cluster.routing.rebalance.enable`) in order to remain in a
balanced state. For example, if the requested allocation includes moving a
shard from `node1` to `node2` then this may cause a shard to be moved from
`node2` back to `node1` to even things out.

The cluster can be set to disable allocations using the
`cluster.routing.allocation.enable` setting.  If allocations are disabled then
the only allocations that will be performed are explicit ones given using the
`reroute` command, and consequent allocations due to rebalancing.

It is possible to run `reroute` commands in "dry run" mode by using the
`?dry_run` URI query parameter, or by passing `"dry_run": true` in the request
body. This will calculate the result of applying the commands to the current
cluster state, and return the resulting cluster state after the commands (and
re-balancing) has been applied, but will not actually perform the requested
changes.

If the `?explain` URI query parameter is included then a detailed explanation
of why the commands could or could not be executed is included in the response.

The commands supported are:

`move`::
    Move a started shard from one node to another node. Accepts
    `index` and `shard` for index name and shard number, `from_node` for the
    node to move the shard from, and `to_node` for the node to move the
    shard to.

`cancel`::
    Cancel allocation of a shard (or recovery). Accepts `index` and `shard` for
    index name and shard number, and `node` for the node to cancel the shard
    allocation on. This can be used to force resynchronization of existing
    replicas from the primary shard by cancelling them and allowing them to be
    reinitialized through the standard recovery process. By default only
    replica shard allocations can be cancelled. If it is necessary to cancel
    the allocation of a primary shard then the `allow_primary` flag must also
    be included in the request.

`allocate_replica`::
    Allocate an unassigned replica shard to a node. Accepts `index` and `shard`
    for index name and shard number, and `node` to allocate the shard to. Takes
    <<modules-cluster,allocation deciders>> into account.

[float]
=== Retrying failed allocations

The cluster will attempt to allocate a shard a maximum of
`index.allocation.max_retries` times in a row (defaults to `5`), before giving
up and leaving the shard unallocated. This scenario can be caused by
structural problems such as having an analyzer which refers to a stopwords
file which doesn't exist on all nodes.

Once the problem has been corrected, allocation can be manually retried by
calling the <<cluster-reroute,`reroute`>> API with the `?retry_failed` URI
query parameter, which will attempt a single retry round for these shards.

[float]
=== Forced allocation on unrecoverable errors

Two more commands are available that allow the allocation of a primary shard to
a node. These commands should however be used with extreme care, as primary
shard allocation is usually fully automatically handled by Elasticsearch.
Reasons why a primary shard cannot be automatically allocated include the
following:

- A new index was created but there is no node which satisfies the allocation
  deciders.
- An up-to-date shard copy of the data cannot be found on the current data
  nodes in the cluster. To prevent data loss, the system does not automatically
promote a stale shard copy to primary.

The following two commands are dangerous and may result in data loss. They are
meant to be used in cases where the original data can not be recovered and the
cluster administrator accepts the loss. If you have suffered a temporary issue
that can be fixed, please see the `retry_failed` flag described above. To
emphasise: if these commands are performed and then a node joins the cluster
that holds a copy of the affected shard then the copy on the newly-joined node
will be deleted or overwritten.

`allocate_stale_primary`::
    Allocate a primary shard to a node that holds a stale copy. Accepts the
    `index` and `shard` for index name and shard number, and `node` to allocate
    the shard to. Using this command may lead to data loss for the provided
    shard id. If a node which has the good copy of the data rejoins the cluster
    later on, that data will be deleted or overwritten with the data of the
    stale copy that was forcefully allocated with this command. To ensure that
    these implications are well-understood, this command requires the flag
    `accept_data_loss` to be explicitly set to `true`.

`allocate_empty_primary`::
    Allocate an empty primary shard to a node. Accepts the `index` and `shard`
    for index name and shard number, and `node` to allocate the shard to. Using
    this command leads to a complete loss of all data that was indexed into
    this shard, if it was previously started. If a node which has a copy of the
    data rejoins the cluster later on, that data will be deleted. To ensure
    that these implications are well-understood, this command requires the flag
    `accept_data_loss` to be explicitly set to `true`.