[
https://issues.apache.org/jira/browse/SOLR-12827?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16638029#comment-16638029
]
Shalin Shekhar Mangar commented on SOLR-12827:
----------------------------------------------
This patch migrates the nested cluster properties format to:
{code}
{
"defaults" : {
"collection": {
"numShards": 2,
"numNrtReplicas": 2,
}
}
}
{code}
The old format can be read from ZK. It can also be written using the APIs.
However, it is auto-converted to the new format on reads and writes. We can get
rid of the conversion logic in Solr 9.
The patch also updates the reference guide to mention only the new format. I
added a note saying that the old API syntax continues to be supported but it
will be auto-converted.
I'll commit once tests and precommit pass.
> Migrate cluster wide defaults syntax in cluster properties to a nested
> structure
> --------------------------------------------------------------------------------
>
> Key: SOLR-12827
> URL: https://issues.apache.org/jira/browse/SOLR-12827
> Project: Solr
> Issue Type: Task
> Security Level: Public(Default Security Level. Issues are Public)
> Components: SolrCloud
> Reporter: Shalin Shekhar Mangar
> Priority: Major
> Fix For: 7.6, master (8.0)
>
> Attachments: SOLR-12827.patch
>
>
> SOLR-12387 introduced cluster wide defaults for collection properties.
> However, the syntax committed is flat and ugly e.g. collectionDefaults
> instead of defaults/collection inspite of the latter being proposed by
> Andrzej. Anyway, we should change the format now before it sees widespread
> use. I propose to fix documentation to describe the new format and
> auto-convert cluster properties to new syntax on startup.
--
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
