Thread: BUG #12826: Streaming replication - should master/slave file system be identical?

The following bug has been logged on the website:

Bug reference:      12826
Logged by:          Jamie Koceniak
Email address:      jkoceniak@mediamath.com
PostgreSQL version: 9.1.14
Operating system:   Linux
Description:

Should the file system between a PG master and slave be identical?
When I do an audit of the PG file system (i.e.
postgresql/9.1/main/base/31858), I see _fsm files that exist on the slave
but don’t exist on the master. I also see _fsm & _vm files on the master
that are not on the slave.

Besides doing basic table row counts between a master and slave, how do you
verify that the master and slave are in synch when using streaming
replication?

On Tue, Mar  3, 2015 at 10:19:32PM +0000, jkoceniak@mediamath.com wrote:
> The following bug has been logged on the website:
>
> Bug reference:      12826
> Logged by:          Jamie Koceniak
> Email address:      jkoceniak@mediamath.com
> PostgreSQL version: 9.1.14
> Operating system:   Linux
> Description:
>
> Should the file system between a PG master and slave be identical?
> When I do an audit of the PG file system (i.e.
> postgresql/9.1/main/base/31858), I see _fsm files that exist on the slave
> but don’t exist on the master. I also see _fsm & _vm files on the master
> that are not on the slave.
>
> Besides doing basic table row counts between a master and slave, how do you
> verify that the master and slave are in synch when using streaming
> replication?

fsm and vm files are created by vacuum and might differ between master
and slaves.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + Everyone has their own god. +