On Wed, Feb 17, 2010 at 04:14:35PM +0100, Nikola Ciprich wrote:
> Well,
> I did a lot of stress tests, and it seems to be pretty random, and what's
> worse, I think it is more then one problem, oh great...
> Seems that at least one of my NFS problems was in fact caused by ext4 regression
> (what FS are you using?)
> applying this one: http://bugzilla.kernel.org/attachment.cgi?id=24975
> improved things for me a lot (see http://bugzilla.kernel.org/show_bug.cgi?id=15231 for
> full discussion).
> I'll watch the machine for next few days and report...
Is there *any* kernel (however old) on which you are confident you can
*not* reproduce the kref bug?
--b.
> regards
> nik
>
> On Tue, Feb 16, 2010 at 03:29:47PM -0500, J. Bruce Fields wrote:
> > On Tue, Feb 16, 2010 at 09:27:20PM +0100, Nikola Ciprich wrote:
> > > > Nothing yet. I'm taking another look.
> > > >
> > > > Do you know what the most recent version was where you did *not* see
> > > > these warnings?
> > > well, actually I decided I'll bisect this one no matter what, so hopefully
> > > we'll see soon :)
> > > I'll report
> >
> > That would be fantastic, thanks. What test-case are you using?
> >
> > Any partial information ("xxxx good, xxxx bad") would also be useful.
> >
> > --b.
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> >
>
> --
> -------------------------------------
> Ing. Nikola CIPRICH
> LinuxBox.cz, s.r.o.
> 28. rijna 168, 709 01 Ostrava
>
> tel.: +420 596 603 142
> fax: +420 596 621 273
> mobil: +420 777 093 799
> http://www.linuxbox.cz
>
> mobil servis: +420 737 238 656
> email servis: servis-Jp3n8lUXroTtwjQa/[email protected]
> -------------------------------------
It was working pretty well when we were using 2.6.27.x, but after upgrade, I
also changed FS from XFS to ext4, (so I can't easily test with 2.6.27 now) and
since then NFS is almost unusable on this machine.
But 2.6.27 is pretty old, so there's been a lot of changes.
But greg released 2.6.32.9-rc1 which contains bunch of NFS-related fixes,
so I'll give it a try and report.
nik
On Sat, Feb 20, 2010 at 06:51:05PM -0500, J. Bruce Fields wrote:
> On Wed, Feb 17, 2010 at 04:14:35PM +0100, Nikola Ciprich wrote:
> > Well,
> > I did a lot of stress tests, and it seems to be pretty random, and what's
> > worse, I think it is more then one problem, oh great...
> > Seems that at least one of my NFS problems was in fact caused by ext4 regression
> > (what FS are you using?)
> > applying this one: http://bugzilla.kernel.org/attachment.cgi?id=24975
> > improved things for me a lot (see http://bugzilla.kernel.org/show_bug.cgi?id=15231 for
> > full discussion).
> > I'll watch the machine for next few days and report...
>
> Is there *any* kernel (however old) on which you are confident you can
> *not* reproduce the kref bug?
>
> --b.
>
> > regards
> > nik
> >
> > On Tue, Feb 16, 2010 at 03:29:47PM -0500, J. Bruce Fields wrote:
> > > On Tue, Feb 16, 2010 at 09:27:20PM +0100, Nikola Ciprich wrote:
> > > > > Nothing yet. I'm taking another look.
> > > > >
> > > > > Do you know what the most recent version was where you did *not* see
> > > > > these warnings?
> > > > well, actually I decided I'll bisect this one no matter what, so hopefully
> > > > we'll see soon :)
> > > > I'll report
> > >
> > > That would be fantastic, thanks. What test-case are you using?
> > >
> > > Any partial information ("xxxx good, xxxx bad") would also be useful.
> > >
> > > --b.
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> > > the body of a message to [email protected]
> > > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > >
> >
> > --
> > -------------------------------------
> > Ing. Nikola CIPRICH
> > LinuxBox.cz, s.r.o.
> > 28. rijna 168, 709 01 Ostrava
> >
> > tel.: +420 596 603 142
> > fax: +420 596 621 273
> > mobil: +420 777 093 799
> > http://www.linuxbox.cz
> >
> > mobil servis: +420 737 238 656
> > email servis: [email protected]
> > -------------------------------------
>
--
-------------------------------------
Ing. Nikola CIPRICH
LinuxBox.cz, s.r.o.
28. rijna 168, 709 01 Ostrava
tel.: +420 596 603 142
fax: +420 596 621 273
mobil: +420 777 093 799
http://www.linuxbox.cz
mobil servis: +420 737 238 656
email servis: [email protected]
-------------------------------------
J. Bruce Fields wrote:
> On Wed, Feb 17, 2010 at 04:14:35PM +0100, Nikola Ciprich wrote:
>> Well,
>> I did a lot of stress tests, and it seems to be pretty random, and what's
>> worse, I think it is more then one problem, oh great...
>> Seems that at least one of my NFS problems was in fact caused by ext4 regression
>> (what FS are you using?)
>> applying this one: http://bugzilla.kernel.org/attachment.cgi?id=24975
>> improved things for me a lot (see http://bugzilla.kernel.org/show_bug.cgi?id=15231 for
>> full discussion).
>> I'll watch the machine for next few days and report...
>
> Is there *any* kernel (however old) on which you are confident you can
> *not* reproduce the kref bug?
I have got a bug that looks a lot like this on 2.6.32.8 yesterday less
than 24 hours after booting the first 2.6.32 kernel but we have been
running on 2.6.31 for 3 months before, so I'm fairly confident that it
is a .31 -> .32 regression.
http://lkml.org/lkml/2010/2/23/70
Jesper
--
Jesper