Fri May 11 15:50:12 PDT 2007
- Previous message: [Slony1-general] How to reduce the time data appears in the slave?
- Next message: [Slony1-general] How to reduce the time data appears in the slave?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hello. I read that Slony uses LISTEN/NOTIFY pgsql features to implement immediate spreading of all the master changes to subsribers: http://www.postgresql.org/docs/8.2/interactive/sql-notify.html But seems information is spreaded not so quickly as I expect. E.g., when I post something to the master, changes may appear on slave with 2-3 seconds delay. I could decrease sync_interval to 100 ms (for example) and it helps, BUT in this case slon daemon from the slave begins to bomb the master with the query: select last_value from "_moikrug_cluster".sl_log_status (10 queries per seconds). It is not good for master's CPU to process such large number of queries (if I have 10 slaves - 100 queries/s, aargh!). So, the question is: how to make the Slony to update slaves as fast as possible without bombing the master with periodical queries? And - what is the detailed meaning of sync_interval? P.S. I tried to watch postgres logs and noticed that slave runs LISTEN/UNLISTEN periodically, but the number of UNLISTEN is greater than the number of LISTEN. So - maybe my configuration is simply broken, and slon does not listen anything at all? Please say what debug information should I post here. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.slony.info/pipermail/slony1-general/attachments/20070512/= 1c90cb31/attachment.htm
- Previous message: [Slony1-general] How to reduce the time data appears in the slave?
- Next message: [Slony1-general] How to reduce the time data appears in the slave?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list