Your schema won’t tell us much at all about indexing speed, it’s the size of the documents that’s the bigger factor.
Yes, you can increase the timeout, see: https://lucene.apache.org/solr/guide/6_6/format-of-solr-xml.html#format-of-solr-xml, distribUpdateConnTimeout and distribUpdateSoTimeout. However, this is really a shot in the dark. You should see something in your logs about “leader initiated recovery” if this is really the problem, there’s no use in fiddling with these settings unless you’re sure this is what’s happening. The frist place I’d start is looking more thoroughly at the logs to see if you can find out _why_ your replica is going into recovery, there should be information. Look in both the leader and replica logs. Best, Erick > On Jul 22, 2020, at 1:32 AM, vishal patel <vishalpatel200...@outlook.com> > wrote: > > Thanks for reply. > >>> The recovery is probably _caused_ by the node not responding to the update > request due to a timeout > Can we increase update request timeout? > >>> What kind of documents are you indexing? I have seen situations where >>> massive > documents take so long that the request times out and starts this process. > Normal document here is my schema file : > https://drive.google.com/file/d/12SHl3YGP7jT4goikBkeyB2s1NX5_C2gz/view > >>> The logs you posted don’t show the reason the node went into recovery > in the first place, that’s the thing I’d concentrate on finding. > Really, we cannot find from the log that why replica goes into recovery > before recovery log there were our insert and search requests. > > Regards, > Vishal Patel > > Sent from Outlook<http://aka.ms/weboutlook> > ________________________________ > From: Erick Erickson <erickerick...@gmail.com> > Sent: Tuesday, July 21, 2020 6:36 PM > To: solr-user@lucene.apache.org <solr-user@lucene.apache.org> > Subject: Re: Replica goes into recovery mode in Solr 6.1.0 > > The recovery is probably _caused_ by the node not responding to the update > request due to a timeout. The JIRA you reference is unrelated I’d guess. > > What kind of documents are you indexing? I have seen situations where massive > documents take so long that the request times out and starts this process. > > The logs you posted don’t show the reason the node went into recovery > in the first place, that’s the thing I’d concentrate on finding. > > Best, > Erick > >> On Jul 21, 2020, at 1:37 AM, vishal patel <vishalpatel200...@outlook.com> >> wrote: >> >> I am using Solr version 6.1.0, Java 8 version and G1GC on production. We >> have 2 shards and each shard has 1 replica. >> Some times my replica goes into recovery mode and when I check my GC log, I >> can not find the GC pause time more than 600 milliseconds. sometimes GC >> pause time goes near to 1 seconds but at that time replica does not go into >> recovery mode. >> >> My Error Log: >> shard: https://drive.google.com/file/d/1F8Bn7jSXspe2HRelh_vJjKy9DsTRl9h0/view >> replica: >> https://drive.google.com/file/d/1y0fC_n5u3MBMQbXrvxtqaD8vBBXDLR6I/view >> >> When I searched my error "org.apache.http.NoHttpResponseException: failed >> to respond" in Google, I found the one Solr jira case : >> https://issues.apache.org/jira/browse/SOLR-7483 >> >> Any one gives me details about that jira case? is it resolved in other jira >> case? >> >> Regards, >> Vishal patel >> >> >> >> >> Sent from Outlook<http://aka.ms/weboutlook> >
