Re: Automatically force zero_damaged_pages while InRecovery? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Automatically force zero_damaged_pages while InRecovery?
Date
Msg-id 8825.1070135713@sss.pgh.pa.us
Whole thread Raw
In response to Re: Automatically force zero_damaged_pages while InRecovery?  (Alvaro Herrera <alvherre@dcc.uchile.cl>)
List pgsql-hackers
Alvaro Herrera <alvherre@dcc.uchile.cl> writes:
> Yes, but is the error message going to stay verbatim?  Maybe the
> category could be decreased ...

The patch I proposed would still cause a WARNING to come out in the
postmaster log.  We could perhaps reduce it to a NOTICE if InRecovery
is true, or have InRecovery add a DETAIL pointing out that this should
be no problem (though the latter behavior couldn't be back-ported to
7.3 very readily).  We'd not want to change the message wording in the
7.4 branch either.

I think this is all a bit academic though, compared to the basic point
of not failing ;-).  In practice, any recovery from WAL is going to
produce a bunch of scary-looking messages in the postmaster log.
We're not likely to remove those messages, since they provide
potentially critical debugging info if it fails to work.  I'm not that
concerned about making them less scary-looking ...
        regards, tom lane


pgsql-hackers by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: Automatically force zero_damaged_pages while InRecovery?
Next
From: "Marc G. Fournier"
Date:
Subject: Re: $Id$ -> $PostgreSQL$ Change