Replica node

Hi,

Is it possible to configure the ES so that a single node will always get
the replica shards assigned? When I was in a single-node configuration I
just upped the number of replicas for every index to 1 and brought a
data-only node to the cluster and of course the replicas all got created on
that single node.

But since the indexes rotate every night today I discovered that the next
index has it's primary shards on this second server which is not exactly
what I'd like to see...

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/fb45ad0d-7057-4df3-88f5-2a5a03310d7e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

You can force it using this sort of process -

Though unless you have a good reason, it's best to just let ES do it's own
thing.

Regards,
Mark Walkom

Infrastructure Engineer
Campaign Monitor
email: markw@campaignmonitor.com
web: www.campaignmonitor.com

On 12 June 2014 14:50, Tommi Lätti sty@blosphere.net wrote:

Hi,

Is it possible to configure the ES so that a single node will always get
the replica shards assigned? When I was in a single-node configuration I
just upped the number of replicas for every index to 1 and brought a
data-only node to the cluster and of course the replicas all got created on
that single node.

But since the indexes rotate every night today I discovered that the next
index has it's primary shards on this second server which is not exactly
what I'd like to see...

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/fb45ad0d-7057-4df3-88f5-2a5a03310d7e%40googlegroups.com
https://groups.google.com/d/msgid/elasticsearch/fb45ad0d-7057-4df3-88f5-2a5a03310d7e%40googlegroups.com?utm_medium=email&utm_source=footer
.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAEM624ZNWFzk6%2Bb7tJuxUEa35Ss%2BNiCNgBvkoKaRD8d-%2BTWCSw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

We're having this same problem, and it looks like the
index-modules-allocation doesn't actually provide any mechanism to
accomplish this request.

The primary use-case on our end is that we have a two-tiered cluster with a
set of indexers and a set of search nodes. We'd like to zone primaries to
the indexing zone and use asynchronous replication to the search zone.
Sadly, the shard allocation filtering doesn't seem to provide any
mechanism to directly enforce this constraint.

The best we've been able to come up with so far, is to use a forced
awareness attribute and initially zone a new time bin to the indexing zone.
This causes ES to allocate primaries in the indexing zone with no initial
replicas. After creation, if we then go back and include the search zone
for the index (after the primaries are allocated) - replicas are created
there and everything works as expected.

This strikes me as both a very useful feature to provide fine-grained
hardware control for different operations, and also an extremely hacky
work-around the present limitations.

On Thursday, June 12, 2014 12:58:07 AM UTC-4, Mark Walkom wrote:

You can force it using this sort of process -
Elasticsearch Platform — Find real-time answers at scale | Elastic

Though unless you have a good reason, it's best to just let ES do it's own
thing.

Regards,
Mark Walkom

Infrastructure Engineer
Campaign Monitor
email: ma...@campaignmonitor.com <javascript:>
web: www.campaignmonitor.com

On 12 June 2014 14:50, Tommi Lätti <s...@blosphere.net <javascript:>>
wrote:

Hi,

Is it possible to configure the ES so that a single node will always get
the replica shards assigned? When I was in a single-node configuration I
just upped the number of replicas for every index to 1 and brought a
data-only node to the cluster and of course the replicas all got created on
that single node.

But since the indexes rotate every night today I discovered that the next
index has it's primary shards on this second server which is not exactly
what I'd like to see...

--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearc...@googlegroups.com <javascript:>.
To view this discussion on the web visit
https://groups.google.com/d/msgid/elasticsearch/fb45ad0d-7057-4df3-88f5-2a5a03310d7e%40googlegroups.com
https://groups.google.com/d/msgid/elasticsearch/fb45ad0d-7057-4df3-88f5-2a5a03310d7e%40googlegroups.com?utm_medium=email&utm_source=footer
.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/d6059898-3eac-4d67-9331-ef3e452eea28%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.