On Tue, 1 Jul 2003, Nathan Straz wrote:
> On Tue, Jul 01, 2003 at 12:42:26PM -0700, Sebastien Boving wrote:
> > ftpd writing to an xfs filesystem (compiled from 2.4.20-18.9XFS1.3.0pre2
> > for RH9 src rpm from oss.sgi.com ftp site, but without kdb) crashed,
> > leaving the following in the logs:
>
> Do you have any idea what was going on before the crash? I can try to
> reproduce this, but I don't see a lot to go on.
the box is doing only nfs serving, and traffic is currently close to 0. At
that point though, i was doing a ftp from another host to this machine,
recursively mput'ing data on the xfs filesystem.
it's while nfsd was writing a particular directory (or files therein, i
can't get passed the dir) that it crashed (segfault), and the first
messages is sent appeared in the logs.
i then logged into it and tried some commands on that dir, and got the
other messages. that's all i know so far.
> > dir it was writing to contains a broken entry, another dir. I can't cd
> > into it, du it, ... all end with Segmentation Fault and print out:
>
> Can you try to print out some information with xfs_db? See the man page
> for how to use it. Perhaps one of the developers will offer up some
> recommended commands.
i will play with that, but first i'm making an extra online backup of the
the 300gb that's still available.
> > ==> what's the problem? should 1.3.0pre2 not be used? (i have this fs
> > since only a week).
>
> It's still a pre-release so it should be used on machines that you can
> spare some downtime on. It is not meant for production use.
i see, my bad. 2 questions then:
- is there a RH9 2.4.18+ kernel src rpm with 1.2 available somewhere, or
what's the best way to build it?
- once i get there (recent rh kernel with 1.2.0), can i mount my 1.3.0pre2
created fs with it? or should i mkfs.xfs it before?
|