Thread: Adding Additional Standby
Hello, We have an existing master-slave setup running on 9.0. I was looking into adding a new hot standby into this setup — I haveadded a script that satisfies the requirements of the archive_command, but it isn't clear to me how to add this new standbyinto the mix. Is it as simple as calling pg_start_backup to create a new base backup and then starting postgresqlon the new machine once the base backup has completed? Aren't there syncing issues to consider? My intuition tellsme that the standbys need to be in sync for the archive_command to work (since it simply passes in a filename that isexpected to get copied over). Initially, I was planning on pausing replication on the existing standby, and running replication on the new stanby to catchup to the same xid as the old standby. However, I then discovered that pausing isn't supported on 9.0 (but will be on9.1). Does stopping postgresql on the existing standby have the same effect? Am I doing something unusual here? Is it possibleto do what I am attempting to do? Thank you very much! Andrew
Andrew Hannon <ahannon@fiksu.com> wrote: > Hello, > > We have an existing master-slave setup running on 9.0. I was looking > into adding a new hot standby into this setup — I have added a script > that satisfies the requirements of the archive_command, but it isn't > clear to me how to add this new standby into the mix. Is it as simple > as calling pg_start_backup to create a new base backup and then > starting postgresql on the new machine once the base backup has > completed? Yes, that's all. You should increase max_wal_senders (on the server) to (count of clients) + reserve (maybe 3 or so) And you need a propper recovery.conf on the new client. Andreas -- Really, I'm not out to destroy Microsoft. That will just be a completely unintentional side effect. (Linus Torvalds) "If I was god, I would recompile penguin with --enable-fly." (unknown) Kaufbach, Saxony, Germany, Europe. N 51.05082°, E 13.56889°