Tue Jan 22 08:03:31 PST 2008
- Previous message: [Slony1-bugs] [Slony] vacuum locks all tables?
- Next message: [Slony1-bugs] [Bug 32] New: Once every two to three months Slony fails to commit confirms
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
From my usage, autovacuum is not a problem for slony. There are some vacuum done by slony itself. But more probably if it is autovacuum, then it is probably some xid wraparound ? Keep in mind that even if autovac is disabled, it will work to reset xid : http://www.postgresql.org/docs/8.1/static/maintenance.html#VACUUM-FOR-WRAPAROUND That was for info, I didn't saw your question in your mail :) Martin Fandel a écrit : > Hi, > > i've installed a Slony-Replication with 2 nodes. Last year, the first > node was the primary and the second node the secondary. This year, > I have switchover the replication. So right now the first node is > secondary and the second node primary. > > Since yesterday, the replication was working fine. But right now all > tables are locked while slon runs a vacuum process and i have seen > about 500 client processes which are hanging. A workaround for me was > killing the slony-vacuum process periodically. After killing it, the > Replication works fine and all clients are stop hanging. > > Here are my enviroment-informations: > > - Slony 1.2.9 > > - Primary (Node 2) > - 4 CPU Xeon 3,6 Ghz > - 4GB RAM > - 15k disks RAID10 > - SLES10 > - PostgreSQL 8.2.5 (compiled from source) > > - Secondary (Node 1) > - 2 CPU Xeon 1,8 Ghz > - 4GB RAM > - 10k disks RAID10 > - OpenSuse 10.2 > - PostgreSQL 8.1.5 (suse rpm) > > I've tried this for getting the slony tables disabled for pg_autovacuum: > > insert into pg_catalog.pg_autovacuum ( > vacrelid, > enabled, > vac_base_thresh, > vac_scale_factor, > anl_base_thresh, > anl_scale_factor, > vac_cost_delay, > vac_cost_limit, > freeze_min_age, > freeze_max_age > ) select > oid, > 'f', > '5000', > '0.2', > '10000', > '0.2', > '-1', > '-1', > '0', > '0' > from > pg_catalog.pg_class > where > relnamespace = (select oid from pg_namespace where nspname = '_' || > 'isohost_cluster') and relhasindex; > > Thanks! > > Best regards, > > Martin > > _______________________________________________ > Slony1-bugs mailing list > Slony1-bugs at lists.slony.info > http://lists.slony.info/mailman/listinfo/slony1-bugs > -- Cédric Villemain Administrateur de Base de Données Cel: +33 (0)6 74 15 56 53 http://dalibo.com - http://dalibo.org
- Previous message: [Slony1-bugs] [Slony] vacuum locks all tables?
- Next message: [Slony1-bugs] [Bug 32] New: Once every two to three months Slony fails to commit confirms
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-bugs mailing list