pg_resetxlog for 8.0 (was Re: production server down)

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Joe Conway <mail(at)joeconway(dot)com>
Cc: Michael Fuhr <mike(at)fuhr(dot)org>, "Hackers (PostgreSQL)" <pgsql-hackers(at)postgresql(dot)org>
Subject: pg_resetxlog for 8.0 (was Re: production server down)
Date: 2004-12-19 00:31:21
Message-ID: 7337.1103416281@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Joe Conway <mail(at)joeconway(dot)com> writes:
> The manpage for pg_resetxlog gives some general idea how it is used, and
> a way to estimate the next transaction id and wal segment.

I had forgotten that that text was in there. It needs to be updated for
8.0 because WAL segment file names are now three-parters. I think
there's also a "MUST FIX FOR 8.0" to-do item here: there needs to be
a way to specify the timeline ID to use. In prior versions we just
allowed pg_resetxlog to set startup ID to 1 all the time, because it
wasn't really being used for anything. But as of 8.0 that's a
significant value and so you'd better be able to set it.

What I'm inclined to do is just widen the "-l" option to take three
numbers instead of 2. That will keep it in sync with what people will
see when they look at WAL file names. Any objections?

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message David Walker 2004-12-19 00:32:05 Re: Call for porting reports
Previous Message Tom Lane 2004-12-19 00:12:31 Re: production server down