Re: pgsql: Report which WAL sync method we are trying to change *to* when it - Mailing list pgsql-committers

From Magnus Hagander
Subject Re: pgsql: Report which WAL sync method we are trying to change *to* when it
Date
Msg-id 20080512210630.050875a7@mha-laptop.hagander.net
Whole thread Raw
In response to Re: pgsql: Report which WAL sync method we are trying to change *to* when it  (Simon Riggs <simon@2ndquadrant.com>)
Responses Re: pgsql: Report which WAL sync method we are trying to change *to* when it
List pgsql-committers
Simon Riggs wrote:
> On Mon, 2008-05-12 at 14:27 +0000, Magnus Hagander wrote:
> > Log Message:
> > -----------
> > Report which WAL sync method we are trying to change *to* when it
> > fails, not which one we had before (that worked, and thus is
> > completley irrelevant)
>
> Interesting perspective.
>
> If it breaks, I'd rather be able to put it back the way it was than
> regret in technicolour that my new choice was a bad one. ;-)

Well, the message itself indicated that it was the new one...


> Not everybody keeps a change log.
>
> Could we report both?

Yes, we could easily do that if we want to.

But - this is not the error you get when you try to set it. It's the
error you get when you try to *use* it. And really, it's a "should
never happen" error. (The reason it happens this time is due to another
bug). So I don't think doing so would actually help your case - it's
already covered elsewhere in the code where we'll rollback the setting
when you try to change it instead of PANICing.

//Magnus

pgsql-committers by date:

Previous
From: Simon Riggs
Date:
Subject: Re: pgsql: Report which WAL sync method we are trying to change *to* when it
Next
From: Tom Lane
Date:
Subject: Re: pgsql: Report which WAL sync method we are trying to change *to* when it