We are using SuSE Linux 7.2, with Kernel 2.4.14 (SMP) with the official
1.0.2 XFS patches, compiled with gcc 2.95.3. Machine is an IBM X240 with
ServeRaid 4 adapter. It has two CPUs and 2 GB ram.
We had problems when writing with multiple processes to a single file,
(with the 1.0.1 Release), but that is history now.
Since we did the update, the system had been running fine for 70 days, a
pretty heavily loaded webserver with a database (on block devices)
A few days ago, I gzipped the logs (files with size of about 1 GB), and
processes started to be in "D" status. They piled up and when the load was
40, I rebooted the system the hard way (reboot -d -f -n). The system was
working fine (except for everything in /var).
Neither dmesg, nor system logs said anything.
The xfs filesystem mounted on /var (about 9 GB) could not be mounted, the
system just hanged while starting up. I started with runlevel s, did
xfs_repair and everything worked okay.
Now I am trying to reproduce the problem on a similar machine, but I
thought you might help as well.
Thanks for a great filesystem, except for this failure, we had very good
experiences with xfs.
Mit freundlichen Grüßen / With regards,
RZNet AG, Kerpen, Germany
>>>>>>>>>>>>>>>Neue Telefon Nummer: >>>>>>>>Tel:02273-603172