Hi,
I have a elasticsearch cluster (0.19.11) consisting of 4 nodes on amazon
ec2 with replica level set to 3.
Now i've added a 5. node, and the cluster went immediately into yellow
state.
The reason is that cluster has a unassigned _percolator shard.
In the log of the new node i can see:
[2012-12-21 16:34:41,065][DEBUG][indices.store ] [elastic40]
[_percolator][0] deleting shard that is no longer used
but why is that? my understanding is that the _percolator shard should be
on each node? Or am i missing something?
another maybe related issue is that had allocation awareness settings for
my cluster set and its unclear to me how to remove these settings.
e.g.
"cluster.routing.allocation.awareness.attributes" : "datacenter",
"cluster.routing.allocation.awareness.force.datacenter.values" :
"us_east_1a_ec2,us_east_1b_ec2,us_east_1c_ec2,us_east_1d_ec2",
The documentation explains how go get und set these settings, but now how
to delete.
I've set these two values to "" in an attempt to delete them, but shoul the
allocation force option still effective, it might maybe explain the
unallocated _percolator shard.
What is the right way to delete these _cluster settings?
Regards,
Toni
--