[
https://issues.apache.org/jira/browse/SOLR-8083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14910145#comment-14910145
]
Upayavira commented on SOLR-8083:
---------------------------------
Yes please!!
The request I have is that this abstract out functions that Zookeeper offers,
so:
* please tell me the list of live nodes
* please tell me the state of this collection
We get real craziness when the UI depends upon clusterstate.json, which is
replaced by state.json, and then (I assume) the ZookeeperServlet has to rebuild
a clusterstate.json file, pretending that it exists in its file system, to keep
the UI happy.
For the tree view, we still need the ability to list the contents of nodes, but
I'd rather we only ever use that for the tree view.
I can look through the UI source, and give you a list of the API endpoints I
would love to have. I doubt they will be many or hard to implement. Would that
work?
> Convert the ZookeeperInfoServlet to a handler at /admin/zookeeper
> ------------------------------------------------------------------
>
> Key: SOLR-8083
> URL: https://issues.apache.org/jira/browse/SOLR-8083
> Project: Solr
> Issue Type: Improvement
> Reporter: Noble Paul
> Assignee: Noble Paul
> Attachments: SOLR-8083.patch
>
>
> There is no need to have an extra servlet for this purpose
> By keeping this outside of Solr we cannot even secure it properly using the
> security framework and now we have taken up a top level name called zookeeper
> which cannot be used as a collection name
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
