From: | Joachim Wieland <joe(at)mcknight(dot)de> |
---|---|
To: | Fujii Masao <masao(dot)fujii(at)gmail(dot)com> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Hot Standby: too many KnownAssignedXids |
Date: | 2010-11-22 20:49:09 |
Message-ID: | AANLkTinSKsm-vqPTwKjDMoZ7=Jn1OXCz0M1MohEfoHgA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Sun, Nov 21, 2010 at 11:48 PM, Fujii Masao <masao(dot)fujii(at)gmail(dot)com> wrote:
> --------------
> If you suspect a bug in Hot Standby, please set
> trace_recovery_messages = DEBUG2
> in postgresql.conf and repeat the action
>
> Always useful to know
> * max_connections
> * current number of sessions
> * whether we have two phase commits happening
> --------------
The trace_recovery_messages parameter does not give more output...
max_connections is set to 100
there have been no sessions on the standby itself, but a few on the
primary database, I don't know how much but probably not more than 10.
The sessions there were doing quite some load however, among them
slony synchronization, since the hot standby master database was
actually a slony replica.
max_prepared_transactions has not been changed from its default value of 0.
Joachim
From | Date | Subject | |
---|---|---|---|
Next Message | Dimitri Fontaine | 2010-11-22 21:12:39 | Re: Extensions, this time with a patch |
Previous Message | Heikki Linnakangas | 2010-11-22 20:44:10 | Re: directory archive format for pg_dump |