On 05/05/2014 04:19 PM, Michael Renner wrote:
> Hi,
>
> apparently a few users were puzzled that archive_command is ignored
> on slave servers, which comes as a surprise since streaming
> replication will work fine from slaves and as far as I’ve checked the
> documentation also doesn’t point out the fact that archive_command
> gets a different treatment.
>
> Is this intentional or an oversight? Should this be fixed in the code
> (feature parity to SR) or in the documentation making this more
> explicit?
It was intentional, although I can certainly understand the viewpoint
that archive_command should also archive in the standby. IIRC people
argued it both ways when the cascading replication was discussed
The current assumption is that the archive is shared by the master and
standby (or standbys), so that there is no point in archiving the same
file again in the standby. On the contrary, re-archiving the same file
would fail, so you would need to disable archiving in the standby, and
re-enable it when promoting, which would be more complicated.
- Heikki