Обсуждение: replication for hot-standby?
I need something that is able to replicate schema changes as well, so Slony seems out of the game :/ pg_cluster? mammoth postgresql? db_mirror? What to use for a hot-standby box? Thanks in advance
Hi, maybe filesystem-snapshots with wal's are your solution. There is a block-device replication-tool too. It's called drbd (http://www.drbd.org). I never tested/used it, but i heard it should be fine. I'm using rsync/xfs_freeze to make incremental filesystem-snapshots (incremental by hardlinks). This works fine/reliable for me :). best regards, Martin Am Freitag, den 01.07.2005, 15:51 +0200 schrieb Hannes Dorbath: > I need something that is able to replicate schema changes as well, so > Slony seems out of the game :/ > > pg_cluster? > mammoth postgresql? > db_mirror? > > What to use for a hot-standby box? > > Thanks in advance > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster
Hannes Dorbath wrote: > I need something that is able to replicate schema changes as well, so > Slony seems out of the game :/ > > pg_cluster? > mammoth postgresql? > db_mirror? > > What to use for a hot-standby box? > > Thanks in advance > Slony handles schema changes: http://gborg.postgresql.org/project/slony1/genpage.php?slonik_commands#stmt_ddl_script -- Brad Nicholson 416-673-4106 Database Administrator, Afilias Canada Corp.
Yes, it does in /some way/, but it's quite hackish and error prone.. :/ > If there is anything broken about the script, or about how it executes on a particular node, this will cause the slon daemonfor that node to panic and crash. If you restart the node, it will, more likely than not, try to repeat the DDL script,which will, almost certainly, fail the second time just as it did the first time. > Unfortunately, this nonetheless implies that the use of the DDL facility is somewhat fragile and fairly dangerous. > If your applications do not have fairly stable SQL schemas, then using Slony-I for replication is likely to be fraughtwith trouble and frustration. Sorry, this really doesn't sound like something I'd like to use in a production environment.. On 04.07.2005 21:37, Brad Nicholson wrote: > Slony handles schema changes: > http://gborg.postgresql.org/project/slony1/genpage.php?slonik_commands#stmt_ddl_script >
Hannes Dorbath wrote: > Yes, it does in /some way/, but it's quite hackish and error prone.. :/ It's far from hackish, and only error prone if you don't follow best practices (ie:fireing DDL changes through slonik without testing them on _all_ nodes first). > >> If there is anything broken about the script, or about how it >> executes on a particular node, this will cause the slon daemon for >> that node to panic and crash. If you restart the node, it will, more >> likely than not, try to repeat the DDL script, which will, almost >> certainly, fail the second time just as it did the first time. > > >> Unfortunately, this nonetheless implies that the use of the DDL >> facility is somewhat fragile and fairly dangerous. > > >> If your applications do not have fairly stable SQL schemas, then >> using Slony-I for replication is likely to be fraught with trouble >> and frustration. > > > Sorry, this really doesn't sound like something I'd like to use in a > production environment.. > It totally depends on your enviornment. We use Slony in a 24x7x365 production enviornment. Usage requires that you know what your doing, and fair amount of care and planning when using it. But Slony was never designed with the intent of being a click and forget type system. If you have environement where the schema is not stable, replicating with Slony will likely take a lot of work, but can be done safely. If you don't have control over your schema changes (ie a web hosting environement where the end user can change the schema), it's certainly out of the picture. In that scenario, I think you're going to have a hard time finding a replication engine that will let you have a hot standyby. -- Brad Nicholson 416-673-4106 Database Administrator, Afilias Canada Corp.
Brad Nicholson wrote: > It totally depends on your enviornment. We use Slony in a 24x7x365 > production enviornment. Usage requires that you know what your doing, > and fair amount of care and planning when using it. But Slony was > never designed with the intent of being a click and forget type system. > > If you have environement where the schema is not stable, replicating > with Slony will likely take a lot of work, but can be done safely. If > you don't have control over your schema changes (ie a web hosting > environement where the end user can change the schema), it's certainly > out of the picture. In that scenario, I think you're going to have a > hard time finding a replication engine that will let you have a hot > standyby. Seems such a case should use xlog shipping and loading into a hot standby, which we don't support yet. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073