On Wednesday 21 June 2006 4:54 pm, Martin Steigerwald wrote:
> Am Mittwoch 21 Juni 2006 04:18 schrieb Timothy Shimmin:
> > On Wednesday 21 June 2006 5:44 am, Martin Steigerwald wrote:
> > > Am Dienstag 20 Juni 2006 17:58 schrieb ViNiL:
> > > > Can anyone explain to me, what (and why is that :-) is going on
> > > > with the filesystem?
> > >
> > > do you use write caching? With kernel 2.6.16 XFS got corrupted three
> > > times in one week here with writecache enabled.
> >
> > FYI
> > Have you looked at:
> > http://oss.sgi.com/projects/xfs/faq.html#wcache
>
> Hello,
Hi. My procmail didn't like you and I've just seen your email.
>
> yes I did, but still I do not understand the complete picture:
>
> Firstly hdparm -i /dev/somedevice does not show the *current* state of the
> write cache but only the harddisk default state. This can easily be
> verified by using hdparm -W0 and then hdparm -i again. Unfortunately it
> is not possible to query the write cache status via hdparm -W.
>
But how about using "hdparm -I" like the faq suggests? :)
Uppercase -I seems to work for me.
-i Display the identification info that was obtained from the drive
at boot time, if available....
-I Request identification info directly from the drive, which is
displayed in a new expanded format with considerably more detail
than with the older -i flag.
>
> I am have the oppinion, that XFS should deny write access when it detects
> it cannot be safe. But for that it would be necessary to query the
> current harddisks writecache state and whether write barrier is
> available. It write cache is on and write barrier is off, XFS should
> mount read only IMHO and issuing a big fat warning.
>
Yeah, this subject has been discussed before with the XFS team.
--Tim
|