On 9/7/07, Florian Weimer <fweimer(at)bfk(dot)de> wrote:
> * Scott Marlowe:
> > And there's the issue that with windows / NTFS that when one process
> > opens a file for read, it locks it for all other users. This means
> > that things like virus scanners can cause odd, unpredictable failures
> > of your database.
> I think most of them open the file in shared/backup mode. The only
> lock that is created by that guards deletion and renaming. It can
> still lead to obscure failures, but it's not a wholly-eclusive lock.
Well, there've been a lot of issues with anti-virus and postgresql not
getting along. I wonder if pgsql takes out a stronger lock, and when
it can't get it then the failure happens. Not familiar enough with
windows to do more than speculate.
In response to
pgsql-performance by date
|Next:||From: Matthew Schumacher||Date: 2007-09-07 17:21:10|
|Subject: Re: SAN vs Internal Disks|
|Previous:||From: Tom Lane||Date: 2007-09-07 14:56:47|
|Subject: Re: postgres memory management issues? |