Re: Measurin the lag between a master and a replica

From: David Kerr <dmk(at)mr-paradox(dot)net>
To: Ali Pouya <alipouya2(at)gmail(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: Measurin the lag between a master and a replica
Date: 2012-07-19 21:02:34
Message-ID: 9588A29E-845C-42A0-9E42-51A90A3276E0@mr-paradox.net
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-admin


On Jul 19, 2012, at 8:17 AM, Ali Pouya wrote:

> Hi,
> I am using a master and a replica with PostgreSQL 9.1.3.
> I would like to measure the replication lag between my servers (in time units).
>
> On the replica I can use the function pg_last_xact_replay_timestamp().
> But on the master I only can use pg_current_xlog_location() which does not return a timestamp. I do not know any function returning the timestamp.
>
> Is there any way to get the timestamp relative to the result of pg_current_xlog_location() ?
>
> Thanks for your help
> Best regards
> Ali Pouya
>

I was looking into this last month and some nice folks on the list pointed me to this formula
for calculating lag. (to be run on the slave) similar to what Sergey said.

when pg_last_xlog_receive_location() = pg_last_xlog_replay_location()
then 0
else extract ('epoch' from now()-pg_last_xact_replay_timestamp()) end as log_delay;

In response to

Responses

Browse pgsql-admin by date

  From Date Subject
Next Message Madhu.Lanka 2012-07-20 02:33:36 ERROR: there is no parameter $1
Previous Message Umer Asghar 2012-07-19 16:19:39 Postgres Database got down