On Mon, 2005-04-25 at 13:00 -0400, Chris Penney wrote: > Is is safe to run jfs_fsck -n on a live file system? Does it even > make sense?
Yes, it is safe, but no, it doesn't really make sense. Changes to some parts of the metadata are not synced to disk until the volume is unmounted, and other metadata is not guaranteed to be written to disk at any given time. If you are able to remount the volume read-only, you should get reliable output from jfs_fsck, but I know this may not always be an option. > The reason I ask is beause I had an NFS kernel oops and > jfs was in the stack trace. I was considering doing a read-only fsck > just to make sure there were no file system issues after it. The > message I posted to the NFS list is included below in case it's useful > / helpful. In the stack trace, jfs is causing memory pressure, but it's not certain that iput is trapping on a jfs inode. > > Chris > > ---- fowarded message ---- > > I have a 2.6.11.5 NFS server that serves a directory /s that had light > to moderate activity. I had another file server mounted as /nicfs7 > and was doing a local rsync (moving some data) from /s/path to > /nicfs7/path when (for some unknown reason) it hung. I hit ctrl-c and > then kill -9'ed the two remaining rsync processes. I then unmounted > /nicfs7 (which seemed ok right after). I then saw what I pasted in > below in messages. > > At this point NFS service seems ok; however, one of the nfsd threads > is missing. I specify to start 128 and only 127 are running. Is a > missing nfsd a problem? It's no clear to me why an nfsd did in the > first place either when I was using nfs client -> local disk. > > Chris > > Apr 25 10:51:31 nicfs2 kernel: nfs_statfs: statfs error = 512 > Apr 25 10:53:18 nicfs2 kernel: VFS: Busy inodes after unmount. This seems to indicate a problem with /nicfs7 (as you mentioned that you had just unmounted it). I assume was an nfs mount. > Self-destruct in 5 seconds. Have a nice day... It took a little longer... > Apr 25 10:54:40 nicfs2 kernel: Unable to handle kernel paging request > at virtual address 00200038 > printing eip: > c01786b0 > *pde = 00000000 > Oops: 0000 [#1] > SMP > Modules linked in: nls_utf8 af_packet usbhid nfsd exportfs ipv6 > ohci_hcd e1000 i2c_piix4 sworks_agp agpgart i2c_core evdev usbcore jfs > dm_round_robin dm_multipath dm_mod ext3 jbd qla2300 qla2xxx > scsi_transport_fc ips sd_mod scsi_mod > CPU: 1 > EIP: 0060:[<c01786b0>] Not tainted VLI > EFLAGS: 00010202 (2.6.11.5) > EIP is at clear_inode+0x70/0x150 I would guess that this is a problem with the nfs-mounted volume (nfs client, rather than server), and not related to jfs. -- David Kleikamp IBM Linux Technology Center ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Jfs-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jfs-discussion
