Elasticsearch Log message "deleted shard reason [no longer used]" version 2.1.1


(Gerhard) #1

Hi Community!

We are getting debug log messages on our 4-node test environment under heavy index load:


[2016-01-08 11:28:59,363][DEBUG][indices ] [node4] [index_2016-01-02][10] deleted shard reason [no longer used]
[2016-01-08 11:28:59,363][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][10]] active fully on other nodes)]: took 0s no change in cluster_state
[2016-01-08 11:28:59,363][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2015-12-29][1]] active fully on other nodes)]: execute
[2016-01-08 11:28:59,364][DEBUG][indices ] [node4] [index_2015-12-29][1] deleted shard reason [no longer used]
[2016-01-08 11:28:59,364][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2015-12-29][1]] active fully on other nodes)]: took 0s no change in cluster_state
[2016-01-08 11:28:59,364][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][11]] active fully on other nodes)]: execute
[2016-01-08 11:28:59,365][DEBUG][indices ] [node4] [index_2016-01-02][11] deleted shard reason [no longer used]
[2016-01-08 11:28:59,365][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][11]] active fully on other nodes)]: took 0s no change in cluster_state
[2016-01-08 11:28:59,365][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][9]] active fully on other nodes)]: execute
[2016-01-08 11:28:59,366][DEBUG][indices ] [node4] [index_2016-01-02][9] deleted shard reason [no longer used]
[2016-01-08 11:28:59,366][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][9]] active fully on other nodes)]: took 0s no change in cluster_state
[2016-01-08 11:28:59,366][DEBUG][cluster.service ] [node4] processing [indices_store ([[index_2016-01-02][4]] active fully on other nodes)]: execute

We are creating one index per day, only the index of the current day receives new data. Old indices are open (searchable), but static.

Now, we are wondering what the above messages mean. We are not expecting any "shard activity" especially on static indices. As we didn't find any documentation about "no longer used" shards, I would very much appreciate any input on how to interpet those logs.

We are not experiencing problems, I would just like to understand what is happening on the cluster.

Thanks and best regards,

Gerhard


(Lee Hinman) #2

When a shard is being relocated between nodes, before deleting it from the
source node (where it is being relocated away from) the node reaches out to
check that it is fully active on the target nodes.

These messages are the shard being deleted now that the other node(s) have
confirmed that they have the shard and that it is active. It's likely that your
cluster is just relocating shards between the nodes.


(system) #3