[
https://issues.apache.org/jira/browse/SOLR-6164?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gregory Chanan updated SOLR-6164:
---------------------------------
Attachment: SOLR-6164.patch
Here's a patch. It does the following:
- The ZkIndexSchemaReader now adds the copy fields to the managed schema from
what is in Zk. This is done by clearing the copy field info and recreating it.
That seemed more straightforward than trying to maintain the existing copy
fields and diffing what is in Zk.
- Improves the existing tests in a few ways:
- Instead of a long sleep and then checking, we loop with short sleeps and exit
as soon as we see success.
- Combines the put (add field), post (add field) and copy field test cases into
one. This ensures the commands "play nicely" together and reduces the amount
of code / execution time.
- Adds more testing to the copy fields commands. Randomly decide if the copy
field source and dest will be an existing field, a new field, or a dynamic
field.
> Copy Fields Schema additions are not distributed to other nodes
> ---------------------------------------------------------------
>
> Key: SOLR-6164
> URL: https://issues.apache.org/jira/browse/SOLR-6164
> Project: Solr
> Issue Type: Bug
> Components: SolrCloud
> Reporter: Gregory Chanan
> Attachments: SOLR-6164.patch
>
>
> I took a closer look at the tests added in SOLR-6145, and the copy fields
> aren't being distributed -- it just looks that way b/c we are adding the same
> copy field directive multiple times to random servers and by chance, they'll
> be added to all the servers.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
