Sat Jan 12 17:51:57 PST 2008
- Previous message: [Slony1-general] URGENT: Migration of master node locked my database
- Next message: [Slony1-general] URGENT: Migration of master node locked my database
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Sat, 12 Jan 2008, Craig James wrote: >> You can unlock it with the UNLOCK SET command. You can find it here in the >> docs: >> >> http://slony.info/documentation/stmtunlockset.html > > Thanks, Jeff, you saved my day. I've had it happen to me before as well. Interestingly, after unlocking the set in the past, I've been able to successfully lock and move the set almost immediately afterwards. That was a few revisions of slony back though, so your mileage may vary. >> I'd check the logs to see if there are any clues as to why the move set >> failed to complete. > > Nothing jumps out at me. On the master node, there's nothing at all during > the time the script was running. Just a bunch of "DST DEBUG1 cleanupThread" > events, nothing else. > > On the slave node, there are some messages; maybe the attached logfile will > mean something to you. > It looks like it was successfully syncing prior to you killing it and restarting postgres, but we've only got about a minute's worth of log data here. It would probably be interesting to see immediately after you started the move set. You might need to grep out the uninteresting bits. > One fact I discovered: The clocks on the two machines were badly out of sync, > about 30 minutes. I assume this is a Bad Thing? It's officially listed as a requirement here: http://slony.info/documentation/requirements.html You might check the slon logs for any indication this is what caused your problem. I thought slon did its work based entirely on xxid, but I could be wrong. Perhaps one of the developers will comment. In any event, I'm sure you know the clocks should be in sync as a best practice. > > Another fact: After the migrate-node operation failed, I had to restart > Postgres because one Postgres process was using 100% CPU, and several hundred > other processes were waiting on SELECT statements -- the CPU was rapidly > getting wedged and wasn't far from locking up altogether. (These were > web-site visitors trying to do queries, each one caused another blocked > process.) > > So the "lost connection" message may be when I shut down Postgres. > >> BTW, what version of slony are you using? > > Version 1.2.9, Postgres 8.1.4. You should upgrade both. The postgresql 8.1 branch is up to: 8.1.11, so that puts you 7 revisions back. There are lots of important bug fixes in those revisions. Upgrading to 8.1.11 is relatively painless as it can be done in place. You should probably consider upgrading to 8.2.6 when you have a chance. Slony is up to 1.2.12 and also has lots of important bug fixes in those revisions. -- Jeff Frost, Owner <jeff at frostconsultingllc.com> Frost Consulting, LLC http://www.frostconsultingllc.com/ Phone: 650-780-7908 FAX: 650-649-1954
- Previous message: [Slony1-general] URGENT: Migration of master node locked my database
- Next message: [Slony1-general] URGENT: Migration of master node locked my database
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list