We have a cluster on production with two nodes and 3 indices.
We restarted the two nodes yesterday and we saw in Head that the active
shards are found only in one node (see attached image).
There's a way to spread the active shards between the nodes? or ES is
manage the shards allocating himself?
AFAIK there's no way to allocate the primaries evenly between nodes. But in
most situations, that isn't an issue, because both search and indexing load
will be balanced by default between the two nodes:
each indexed documents is actually indexed both in the primary in the
replica
search requests are round robinned between the primaries and replicas
Are you seeing that the load is uneven between the two nodes?
We have a cluster on production with two nodes and 3 indices.
We restarted the two nodes yesterday and we saw in Head that the active
shards are found only in one node (see attached image).
There's a way to spread the active shards between the nodes? or ES is
manage the shards allocating himself?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.