lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Varun Thacker (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (SOLR-7117) AutoAddReplicas should have a cluster wide property for controlling number of cores hosted on each node
Date Tue, 17 Feb 2015 09:36:12 GMT

     [ https://issues.apache.org/jira/browse/SOLR-7117?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Varun Thacker updated SOLR-7117:
--------------------------------
    Attachment: SOLR-7117.patch

Thanks Mark for reviewing the patch.

bq. There is some class that makes it easy to quickly test what nodes are chosen for failover.

Yeah I saw {{SharedFSAutoReplicaFailoverUtilsTest}} . It's super cool!

New patch which adds a test for this property.

Couple of other changes I made - 

- Modified a live node in {{testGetBestCreateUrlMultipleCollections()}}. I believe it was
intended to be node2 and node1 in the first place.

- In this should the check be only for node3 since node2 is already hosting the replica. Unless
"r2-4" meant node2 or node4

{code}
    result = buildClusterState("csr1R*r2-4sr3r4r5", NODE1, NODE2, NODE3);
    createUrl = OverseerAutoReplicaFailoverThread.getBestCreateUrl(result.reader, result.badReplica,
null);
    assertTrue(createUrl.equals(NODE2_URL) || createUrl.equals(NODE3_URL));
{code}


Should we keep the property name "maxCoresPerNode" or something else?



> AutoAddReplicas should have a cluster wide property for controlling number of cores hosted
on each node
> -------------------------------------------------------------------------------------------------------
>
>                 Key: SOLR-7117
>                 URL: https://issues.apache.org/jira/browse/SOLR-7117
>             Project: Solr
>          Issue Type: Improvement
>            Reporter: Varun Thacker
>            Priority: Minor
>             Fix For: Trunk, 5.1
>
>         Attachments: SOLR-7117.patch, SOLR-7117.patch
>
>
> Currently when finding the best node to host the failed replicas, we respect the maxShardsPerNode
property. This is not an ideal solution as it's a per collection property and we need a cluster
wide property. Also using maxShardsPerNode can lead to unequal distribution of replicas across
nodes.
> We should just let users use the CLUSTERPROP API to set the max number of cores to be
hosted on each node and use that value while picking the node the replica will be hosted on.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message