I was experimenting with the
"cluster.routing.allocation.awareness.attributes" setting. I started with
one data node in the cluster, with allocation awareness attribute set to
"value1". I expect that all the shards will be allocated to this node.
Instead, only primary shards are allocated to this node. All the replicas
are unassigned. Note that I have not set the
"cluster.routing.allocation.awareness.force.zone.values" at all.
I'm using ES 1.1.1. Can anyone confirm? Or did I do something stupid.
I was experimenting with the
"cluster.routing.allocation.awareness.attributes" setting. I started with
one data node in the cluster, with allocation awareness attribute set to
"value1". I expect that all the shards will be allocated to this node.
Instead, only primary shards are allocated to this node. All the replicas
are unassigned. Note that I have not set the
"cluster.routing.allocation.awareness.force.zone.values" at all.
I'm using ES 1.1.1. Can anyone confirm? Or did I do something stupid.
Hmm... I'm reading the same document and came to the opposite conclusion.
My understanding is that if all the nodes have the same attribute, then all shards will be deployed among them. As soon as a node with different
attribute arrives, that's when the "replica will not be deployed to nodes
with the same attribute as that of primary" apply. In my case, I only have
one node, hence one attribute available. So all shards should be deployed.
Thanks.
-Yongtao
On Friday, July 25, 2014 3:56:10 PM UTC-7, Mark Walkom wrote:
On 26 July 2014 03:44, Yongtao You <yongt...@gmail.com <javascript:>>
wrote:
Hi,
I was experimenting with the
"cluster.routing.allocation.awareness.attributes" setting. I started with
one data node in the cluster, with allocation awareness attribute set to
"value1". I expect that all the shards will be allocated to this node.
Instead, only primary shards are allocated to this node. All the replicas
are unassigned. Note that I have not set the
"cluster.routing.allocation.awareness.force.zone.values" at all.
I'm using ES 1.1.1. Can anyone confirm? Or did I do something stupid.
Hmm... I'm reading the same document and came to the opposite conclusion.
My understanding is that if all the nodes have the same attribute, then all shards will be deployed among them. As soon as a node with different
attribute arrives, that's when the "replica will not be deployed to nodes
with the same attribute as that of primary" apply. In my case, I only have
one node, hence one attribute available. So all shards should be deployed.
Thanks.
-Yongtao
On Friday, July 25, 2014 3:56:10 PM UTC-7, Mark Walkom wrote:
Take a look at the example here http://www.elasticsearch.org/
guide/en/elasticsearch/reference/current/modules-cluster.html#allocation-
awareness
Basically a shard and it's replica will never be allocated to the same
instance, which is what you are seeing.
I was experimenting with the "cluster.routing.allocation.awareness.attributes"
setting. I started with one data node in the cluster, with allocation
awareness attribute set to "value1". I expect that all the shards will be
allocated to this node. Instead, only primary shards are allocated to this
node. All the replicas are unassigned. Note that I have not set the
"cluster.routing.allocation.awareness.force.zone.values" at all.
I'm using ES 1.1.1. Can anyone confirm? Or did I do something stupid.
Thanks.
Yongtao
--
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.
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.