This sounds like something that should have some scoping criteria for
security (I haven't looked at code yet). Is it possible that User A be
given the ability to delete replicas for just a gien collection? (I don't
see the collection in the url you provided)

On Tue, May 9, 2023 at 1:19 PM Jason Gerlowski <gerlowsk...@gmail.com>
wrote:

> Hey all,
>
> I was looking at /admin/collections?action=DELETEREPLICA recently and
> was surprised to discover that the API is much more powerful than I'd
> realized.  Not only can it delete a single replica from a single
> shard, but it can also be used to delete 'N' replicas from either a
> single shard or all 'M' shards in the collection.  The
> replica-selection logic lives in
> 'DeleteReplicaCmd.pickReplicasToBeDeleted'.  AFAICT its only
> consideration is to avoid picking the shard-leader for deletion.  It
> doesn't consider the location of replicas or even their type.  (Which
> I guess makes sense, since this was added in 6.3, before Solr had
> different replica types.)  [1] [2]
>
> In the replica-type world we live in today though, this seems a little
> dangerous.  Solr could blithely delete all leadership-capable replicas
> except the current leader. etc.
>
> My question: does this functionality still make sense now that Solr
> has replica types?  Should it be deprecated+removed, or maybe modified
> to be replica-type specific?
>
> Best,
>
> Jason
>
> [1] https://issues.apache.org/jira/browse/SOLR-9319
> [2]
> https://github.com/apache/solr/blob/c99af207c761ec34812ef1cc3054eb2804b7448b/solr/core/src/java/org/apache/solr/cloud/api/collections/DeleteReplicaCmd.java#L183
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@solr.apache.org
> For additional commands, e-mail: dev-h...@solr.apache.org
>
>

-- 
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)

Reply via email to