Search Postgresql Archives

Re: Re: Hot standby problems: consistent state not reached, no connection to master server.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On Tue, 14 Apr 2015 11:59 Ilya Ashchepkov <koctep@xxxxxxxxx> wrote:

On Mon, 13 Apr 2015 12:24:11 -0700
Adrian Klaver <adrian.klaver@xxxxxxxxxxx> wrote:

> On 04/13/2015 11:25 AM, Ilya Ashchepkov wrote:
> > On Mon, 13 Apr 2015 10:06:05 -0700
> > Adrian Klaver <adrian.klaver@xxxxxxxxxxx> wrote:
> >
> >> On 04/13/2015 09:42 AM, Ilya Ashchepkov wrote:
> >>> On Sun, 12 Apr 2015 17:30:44 -0700
> >>> Adrian Klaver <adrian.klaver@xxxxxxxxxxx> wrote:
> >>>
> >>
> >>>>>>
> >>>>>
>
> >>>> If a connection is not being made:
> >>>>
> >>>> 1) Dose user replication have REPLICATION rights?
> >>>> 2) Is the pg_hba.conf on the master set up to allow a connection
> >>>> from the standby for user replication and database replication?
> >>>
> >>> I commented 'restore_command' in recovery.conf and after start
> >>> slave connected to master.
> >>> Then I uncomment it back. Is it possible to have a both, streaming
> >>> connection and restoring from wal files from NFS share?
> >>
> >> Yes:
> >>
> >> http://www.postgresql.org/docs/9.3/interactive/warm-standby.html#STREAMING-REPLICATION
> >>
> >> I wonder if your master is recycling WALs fast enough that the
> >> streaming can't find them and the standby has to go to the archive
> >> instead.
> >>
> >> What is your wal_keep_segments on the master set to?:
> > # select name,setting from pg_settings where name like
> > 'wal_keep_segments'; name        | setting
> > -------------------+---------
> >   wal_keep_segments | 128
> >
> >
> > I run tcpdump -ni eth0 port 5432 on slave and didn't see any packet
> > from slave to master after restart.
>
> Just to be clear:
>
> 1) When you comment out the restore_command the standby connects to
> the master, correct?

Yes.

>
> 2) When you uncomment restore_command you do not see a standby
> connection, correct?

Yes.

>
> So:
>
> 1) When you are changing the restore_command status do you restart
> the standby server?

Yes.

>
> 2) What does  select * from pg_stat_replication show, in either case?
>
> www.postgresql.org/docs/9.3/static/monitoring-stats.html#PG-STAT-REPLICATION-VIEW

0 rows on master
0 rows on slave

>
> 3) I may have missed it, but what is your archive_command on the
> master?

# select name,setting from pg_settings where name like 'archive_command';
      name       |                                  setting
-----------------+----------------------------------------------------------------------------
 archive_command | test ! -f /media/psqlbak/wals/main/%f && cp %p /media/psqlbak/wals/main/%f

>
> >
> >>
> >> http://www.postgresql.org/docs/9.3/interactive/runtime-config-replication.html#RUNTIME-CONFIG-REPLICATION-SENDER
> >>
> >>>
> >>>>
> >>>> Where are the WAL files coming from?
> >>>
> >>> NFS share on master.
>
>



Can you share the cluster log for your standby database and also your primary database? Sorry if you have already shared it and I have missed it. If there is an error in connection there are good chances that some hint about it must be logged. Generally the logs will be placed in pg_log inside your data directory. And log_collector must be set to on.


--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general



[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Postgresql Jobs]     [Postgresql Admin]     [Postgresql Performance]     [Linux Clusters]     [PHP Home]     [PHP on Windows]     [Kernel Newbies]     [PHP Classes]     [PHP Books]     [PHP Databases]     [Postgresql & PHP]     [Yosemite]
  Powered by Linux