From: | Magnus Hagander <magnus(at)hagander(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Bruce Momjian <bruce(at)momjian(dot)us>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Fairly serious bug induced by latest guc enum changes |
Date: | 2008-07-01 19:10:34 |
Message-ID: | 486A812A.4050405@hagander.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Tom Lane wrote:
> Magnus Hagander <magnus(at)hagander(dot)net> writes:
>> Or are you talking about changing the variable "fsync"? If so, doesn't
>> "fsync=off" also change the behavior of other parts of the code, so it's
>> not just WAL, which means it'd be pretty unsafe *anyway* unless you
>> actually "sync" the disks, and not just fsync?
>
> No, because the other uses of it are controlling whether to issue
> fsync() calls dynamically. The use in get_sync_bit is the only one
> that sets persistent state. In fact md.c goes out of its way to ensure
> that changing fsync on the fly behaves as expected.
Not having looked at md.c (I confess...) but don't we have a problem in
case we have closed the file without fsyncing it, and then change the
fsync parameter?
Either way, I see your point, but I doubt it's worth getting upset over.
Funning with fsync=off in the first place is bad, and if it takes you
one WAL segment to "recover", I think that's acceptable...
//Magnus
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2008-07-01 19:11:03 | Re: Location for pgstat.stat |
Previous Message | Tom Lane | 2008-07-01 19:07:21 | Re: Fairly serious bug induced by latest guc enum changes |