Re: Forcing current WAL file to be archived - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: Forcing current WAL file to be archived
Date
Msg-id 200607251557.k6PFvUf17298@momjian.us
Whole thread Raw
In response to Re: Forcing current WAL file to be archived  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Forcing current WAL file to be archived  (Simon Riggs <simon@2ndquadrant.com>)
List pgsql-hackers
Tom Lane wrote:
> Simon Riggs <simon@2ndquadrant.com> writes:
> > I was planning to add a new GUC
> >     archive_timeout (integer) = max # secs between log file switches
> 
> That's fine, but feature freeze is in a week and we don't even have the
> basic function for manually doing a log file switch.  Let's get that
> done first and then think about automatic switches.
> 
> Also, forcing another 16M of WAL out every few seconds is a pretty crude
> and inefficient way of making sure your hot standby server is up to
> date.  As Hannu noted, an archiving script can do better than that if it
> can access the current WAL file name and offset.  So we really want to
> get the function to return that info done as well.

Agreed.  One concern I have is that we have waited for Simon to complete
this for 1.5 years, and now with a week left he is still working on it
(or starting on it).  I am wondering if someone else needs to take this
on, because if Simon doesn't complete it in a week, we don't have it for
8.2, and we would then have to sit through another year of PITR
complaints.  :-(

--  Bruce Momjian   bruce@momjian.us EnterpriseDB    http://www.enterprisedb.com
 + If your life is a hard drive, Christ can be your backup. +


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Forcing current WAL file to be archived
Next
From: Stephen Frost
Date:
Subject: Re: Forcing current WAL file to be archived