Changing cluster name on 2-node cluster causes IndexMissingException

Steps to reproduce:

  1. Create a 2-node cluster
  2. Create a new index
  3. Add one doc to the index
  4. Take cluster down using Nodes Shutdown API
  5. Change cluster.name in elasticsearch.yml on both nodes
  6. service elasticsearch start on both nodes
  7. Cluster comes up OK, looks like working cluster in BigDesk
  8. Queries against your previously-working index return
    IndexMissingException

And if I change the cluster name back to its original name and restart the
cluster, the index works fine again.

Is this a bug? Design limitation? Pilot error?

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

If you want to rename a cluster you will have to modify that name also in /data dir.

For example:

mv ./data/elasticsearch ./data/mynewclustername

On all nodes.

--
David :wink:
Twitter : @dadoonet / @elasticsearchfr / @scrutmydocs

Le 22 mai 2013 à 20:58, Brian Gadoury bgadoury@endpoint.com a écrit :

Steps to reproduce:
Create a 2-node cluster
Create a new index
Add one doc to the index
Take cluster down using Nodes Shutdown API
Change cluster.name in elasticsearch.yml on both nodes
service elasticsearch start on both nodes
Cluster comes up OK, looks like working cluster in BigDesk
Queries against your previously-working index return IndexMissingException
And if I change the cluster name back to its original name and restart the cluster, the index works fine again.

Is this a bug? Design limitation? Pilot error?

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

Brian Gadoury wrote:

[...]

And if I change the cluster name back to its original name and
restart the
cluster, the index works fine again.

Is this a bug? Design limitation? Pilot error?

Feature. The index data reside on disk in a path partially
defined by the cluster name. Once you changed it, ES created a
new folder structure. When you changed it back, it used the
original data.

Drew

--
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.
For more options, visit https://groups.google.com/groups/opt_out.

Thanks gents!

--
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.
For more options, visit https://groups.google.com/groups/opt_out.