Re: transactoin id wraparound problem

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Andrew Sullivan <ajs(at)crankycanuck(dot)ca>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: transactoin id wraparound problem
Date: 2006-09-06 19:42:41
Message-ID: 23650.1157571761@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Andrew Sullivan <ajs(at)crankycanuck(dot)ca> writes:
> On Wed, Sep 06, 2006 at 12:23:01PM -0700, Sriram Dandapani wrote:
>> Is there a way to monitor vacuum progress. Can I resume normal
>> operations assuming vacuum will update the transaction ids or should I
>> wait till it finishes.

> That depends on how many transactions you think will happen while
> vacuum is running. If it's a lot, you could pass the fatal point (==
> you lose data) before vacuum finishes.

He's apparently running 8.1 (I don't think that error message wording
exists in older releases), so what would happen is that the DB will
refuse to start new transactions; but there won't be any data lost.

Still it might be best to let the vacuum have all the machine's cycles
to get done as fast as possible.

regards, tom lane

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Alvaro Herrera 2006-09-06 19:44:34 Re: Why so long between archive calls?
Previous Message Chris Hoover 2006-09-06 19:40:44 Re: Why so long between archive calls?