Re: PeerSyncReplicationTest failure

2017-01-16 Thread Erick Erickson
100 beast runs and no failures so this looks fixed by Alan's latest SOLR-9906 push. On Mon, Jan 16, 2017 at 9:17 AM, Erick Erickson wrote: > This is probably SOLR-9906 right? I'll go start my beasting from > yesterday on a new pull. > > On Sun, Jan 15, 2017 at 8:11 PM,

Re: PeerSyncReplicationTest failure

2017-01-16 Thread Erick Erickson
This is probably SOLR-9906 right? I'll go start my beasting from yesterday on a new pull. On Sun, Jan 15, 2017 at 8:11 PM, Erick Erickson wrote: > Pushkar: > > Yes, PeerSynchReplicationTest. I'm getting 21/100 failures when > beasting on 6x so it's not a trunk-only

Re: PeerSyncReplicationTest failure

2017-01-15 Thread Erick Erickson
Pushkar: Yes, PeerSynchReplicationTest. I'm getting 21/100 failures when beasting on 6x so it's not a trunk-only issue. The script I was using is Mark Miller's "The best Lucene / Solr beasting script in the world. TM." here: https://gist.github.com/markrmiller/dbdb792216dc98b018ad Here's the

Re: PeerSyncReplicationTest failure

2017-01-15 Thread Erick Erickson
Thanks Uwe, I didn't _think_ what I was seeing was possible. I was using Mark's beast script with the option set that's supposed to do a clean-and-compie first when I saw that error, so I don't quite know what was up with that. But with a manual clean that error went away so I'm on to others.

Re: PeerSyncReplicationTest failure

2017-01-15 Thread Pushkar Raste
Erick, Is this PeerSyncTest or PeerSyncReplicationTest? Can you send me link to Jenkins build logs(if this is happening on Jenkins). I recently sent a patch to improve the validation check (in the PeerSyncReplicationTest) made in order to figure out if node successfully recovered via

Re: PeerSyncReplicationTest failure

2017-01-15 Thread Uwe Schindler
Hi, Are you sure that you ran ant clean on top level? It is impossible that this error occurs if compilation was fine. Uwe Am 15. Januar 2017 22:10:27 MEZ schrieb Erick Erickson : >I was wondering about the failures and tried to beast it on my Pro on >trunk which