Обсуждение: BUG #14125: when master set tablespace,slave recovery xlog pending

The following bug has been logged on the website:

Bug reference:      14125
Logged by:          YongWang
Email address:      wangyong20009109@163.com
PostgreSQL version: 9.4.5
Operating system:   Red Hat 4.8.2-16
Description:

when master alter database set tablespace,the slave(async replication)
recovery xlog is pending,the slave normal receiver xlog.

slave ps -ef|grep postgres:
postgres 25636     1  0 Apr26 ?        00:00:09 /pgdir/pg9.4.5/bin/postgres
-D /pgdata
postgres 25741 25636  0 Apr26 ?        00:00:00 postgres: logger process

postgres 25742 25636  6 Apr26 ?        13:30:41 postgres: startup process
recovering 000000010000019E00000035
postgres 25768 25636  0 Apr26 ?        01:45:15 postgres: checkpointer
process
postgres 25769 25636  0 Apr26 ?        00:27:00 postgres: writer process

postgres 25854 25636  0 Apr26 ?        00:01:41 postgres: stats collector
process
postgres 26868 25636  0 Apr26 ?        00:41:54 postgres: wal receiver
process   streaming 1A0/938B17D0

On Wed, May 4, 2016 at 4:26 PM,  <wangyong20009109@163.com> wrote:
> The following bug has been logged on the website:
>
> Bug reference:      14125
> Logged by:          YongWang
> Email address:      wangyong20009109@163.com
> PostgreSQL version: 9.4.5
> Operating system:   Red Hat 4.8.2-16
> Description:

> when master alter database set tablespace,the slave(async replication)
> recovery xlog is pending,the slave normal receiver xlog.

Is the slave complaining about something in its logs? Also, do you
have unlogged relations on this database? We fixed in 9.4.6 a problem
with ALTER TABLE SET TABLESPACE that was not emitting correctly WAL
records. I cannot reproduce the problem to be honest, so you may want
to retry with 9.4.7 first, which is the last released version in the
9.4 series.
--
Michael