Messages in this thread | | | From | Sanjoy Mahajan <> | Subject | directory sread error on 2.0.27 | Date | 27 Mar 1998 12:01:43 -0800 |
| |
On our P5/90 2.0.27 system (RedHat 4.1 mostly), the following command (run at 1am by the cron daemon):
/usr/bin/updatedb --localpaths='' --netpaths='/' --netuser=nobody \ --prunepaths='/tmp /proc /mnt /var/tmp /var/spool /dev /home \ /net /.automount' 2> /dev/null
regularly produces the following errors in the /var/log/messages:
Mar 27 01:03:32 dove kernel: Directory sread (sector 32) failed Mar 27 01:03:32 dove kernel: attempt to access beyond end of device Mar 27 01:03:32 dove kernel: 02:00: rw=0, want=16, limit=2 ...
It's an IDE drive, PCI bus.
Is this a hardware problem, a kernel problem, or an NFS problem? If it's a kernel problem, is it already fixed (in which case I'll upgrade)?
I wonder if it is an NFS problem, since some of the directories that are getting updatedb'ed are across NFS (from SunOS 4.1.4 Sparc-10s). We're running Universal NFS Server 2.2beta16.
It might also be a hardware problem, since we got this computer from what turned out to be a slightly dodgy computer store. On boot, the kernel detects a "buggy RZ1000 inteface" and disables readahead.
Let me know if you need more information. I can't run extensive tests, since this is a production machine (I'm producing my thesis on it, which is due shortly).
With thanks,
/Sanjoy
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |