lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erick Erickson <erickerick...@gmail.com>
Subject RE: Getting "collection already exists" when creating collection in admin UI
Date Fri, 12 Aug 2016 19:24:27 GMT
Reinstalling ZK wouldn't help if the data directory weren't purged..m

On Aug 12, 2016 11:35, "Alexandre Drouin" <alexandre.drouin@orckestra.com>
wrote:

> Thanks for the offer however I think I have a different issue.  I
> reinstalled my ZK and Solr servers between each tests so I didn't have any
> unwanted files.
>
>
> Alexandre Drouin
>
> -----Original Message-----
> From: John Bickerstaff [mailto:john@johnbickerstaff.com]
> Sent: August 12, 2016 1:43 PM
> To: solr-user@lucene.apache.org
> Subject: Re: Getting "collection already exists" when creating collection
> in admin UI
> Importance: High
>
> I don't know if this helps, but I had trouble creating collections due to
> a number of issues and I think I got this error (I was using the command
> line, not the UI)
>
> As I recall, if it exists in Zookeeper, it will error out.  It was a while
> ago, but I think the way I had to solve it was to go into Zookeeper and
> delete the "node".
>
> This was easier for me because I was using "chroot" in Zookeeper such that
> each collection was separate - so all I had to do was delete the entire
> node and start over.
>
> Take me with a grain of salt - it was a while ago.
>
> If you want, I have linux command lines for most / all of this... let me
> know.
>
> On Fri, Aug 12, 2016 at 11:10 AM, Alexandre Drouin <
> alexandre.drouin@orckestra.com> wrote:
>
> > Hi Esther-Melaine,
> >
> > The collection exists in Zookeeper under the /collections node and I
> > can see the shardX_replicaX folders under $SOLR_HOME/server/solr of
> > both servers.
> >
> > I was not able to replicate the issue using the collection API.  Here
> > are the logs where I added the 'MyNewerNode'
> > https://gist.github.com/orck-
> > adrouin/4d074cbb60141cba90c0aae9c55360d4
> >
> > I took a closer look at the admin UI and here are my findings:
> >   - In Chrome's devtool I can see the first create request
> >   - After 10 seconds the request getting aborted and a second create
> > request is sent to the server
> >   - In Fiddler I can see that the first request completes successfully
> > without any issues.  The second request is sent a few seconds before
> > the first one ends so it looks like a admin UI issue.
> >
> > Is it possible that the admin UI has some kind of TTL for requests set
> > to
> > 10 seconds?
> >
> > You mentioned something about the nodes going into recovery.  Any idea
> > how I can fix this issue?
> >
> > My development environment (if it makes a difference):
> >   - OS: Windows
> >   - 2 Solr 6.1 nodes using SolrCloud.  They both are running on the
> > same server using different ports.
> >   - Zookeeper 3.4.8
> >
> > Alexandre Drouin
> >
> >
> > -----Original Message-----
> > From: Esther-Melaine Quansah [mailto:esther.quansah@lucidworks.com]
> > Sent: August 12, 2016 10:46 AM
> > To: solr-user@lucene.apache.org
> > Subject: Re: Getting "collection already exists" when creating
> > collection in admin UI
> > Importance: High
> >
> > Hi Alexandre,
> >
> > The question here is why the create action is called twice. You’re
> > getting that “collection already exists” error after the second action
> is called.
> > Can you verify if MyNewNode exists in /collections in ZK or on the
> > machines running Solr at $SOLR_HOME/server/solr/ Your logs show a lot
> > of issues around the overseer and it looks like those nodes are going
> > into recovery pretty frequently. Can you replicate this issue by
> > creating a collection through the API (not through the UI):
> >
> > http://localhost:8983/admin/collections?action=CREATE&
> > name=MyNewerNode&numShards=1&replicationFactor=2&
> > maxShardsPerNode=1&collection.configName=DefaultConfig
> >
> > Thanks,
> > Esther
> >
> >
> > > On Aug 12, 2016, at 10:05 AM, Alexandre Drouin <
> > alexandre.drouin@orckestra.com> wrote:
> > >
> > > Hello,
> > >
> > > I am running SolrCloud with 2 nodes (Solr 6.1 with SSL and basic
> > > auth)
> > and with one Zookeeper node (for development purposes) and when I try
> > to create a new collection in the admin UI with 'replicationFactor=2'
> > I get a "Connection to Solr lost" message and another message telling me
> "
> > collection already exists: MyNewNode".  I made sure that a collection
> > with the same name does not exists and the issue does not appear with
> > a replication factor of 1.
> > >
> > > While debugging I saw that the create action is called twice with
> > > the following parameters:
> > > /solr/admin/collections?_=1471010473184&action=CREATE&collection.con
> > > fi
> > > gName=DefaultConfig&maxShardsPerNode=1&name=aaa&numShards=1&replicat
> > > io nFactor=2&router.name=compositeId&routerName=compositeId&wt=json
> > >
> > > Can anyone replicate this issue?  I have not found it in JIRA.
> > >
> > >
> > > Below is the relevant log (if useful) and I posted the full logs
> > > here
> > > https://gist.github.com/orck-adrouin/690d485ba0835320273e7b2e09fb377
> > > 1
> > >
> > > 63549 ERROR
> > > (OverseerThreadFactory-5-thread-5-processing-n:orc-dev-solr-cd.local
> > > :8444_solr)
> > [   ] o.a.s.c.OverseerCollectionMessageHandler Collection: MyNewNode
> > operation: create failed:org.apache.solr.common.SolrException:
> > collection already exists: MyNewNode
> > >       at org.apache.solr.cloud.OverseerCollectionMessageHandl
> > er.createCollection(OverseerCollectionMessageHandler.java:1832)
> > >       at org.apache.solr.cloud.OverseerCollectionMessageHandl
> > er.processMessage(OverseerCollectionMessageHandler.java:224)
> > >       at org.apache.solr.cloud.OverseerTaskProcessor$Runner.
> > run(OverseerTaskProcessor.java:463)
> > >       at org.apache.solr.common.util.ExecutorUtil$
> > MDCAwareThreadPoolExecutor.lambda$execute$22(ExecutorUtil.java:229)
> > >       at java.util.concurrent.ThreadPoolExecutor.runWorker(
> > ThreadPoolExecutor.java:1142)
> > >       at java.util.concurrent.ThreadPoolExecutor$Worker.run(
> > ThreadPoolExecutor.java:617)
> > >       at java.lang.Thread.run(Thread.java:745)
> > >
> > > Thanks,
> > > Alexandre Drouin
> >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message