[
https://issues.apache.org/jira/browse/SOLR-12592?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16566364#comment-16566364
]
ASF subversion and git services commented on SOLR-12592:
--------------------------------------------------------
Commit 600c15d14e73274d4152e8ef1b8c0d0aae69fd18 in lucene-solr's branch
refs/heads/branch_7x from [~noble.paul]
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=600c15d ]
SOLR-12592: support #EQUAL function in cores in autoscaling policies
> Support the cores:'EQUAL' in autoscaling policies
> -------------------------------------------------
>
> Key: SOLR-12592
> URL: https://issues.apache.org/jira/browse/SOLR-12592
> Project: Solr
> Issue Type: New Feature
> Security Level: Public(Default Security Level. Issues are Public)
> Components: AutoScaling, SolrCloud
> Reporter: Noble Paul
> Assignee: Noble Paul
> Priority: Major
>
> It's possible to sort nodes according to cores and the system can normally
> prefer nodes with fewer cores when placing new replicas. However, it may not
> give suggestions to move off replicas , if the system is already in an
> unbalanced state.
> The following rule may help achieve such a balanced distribution of cores
> {code}
> {"cores":"#EQUAL" , node: "#ANY"}
> {code}
> The value of cores is computed as {{total_cores/total_nodes}}. for e.g: if
> there are 28 cores in total and there are 5 nodes . the value of cores= 28/5
> = 5.6. This means a node may have either 5 cores or 6 cores.
> It's possible that this may cause a conflict with other collection-specific
> rules such as
> {code}
> {"replica":"#EQUAL" , "node" : "#ANY"}
> {code}
> It can be remedied by making this rule , not strict.
> {code}
> {"cores":"#EQUAL" , "node": "#ANY", "strict"=false}
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
