On Thu, 08 Jun 2006 21:39:45 +0200, prasad <[EMAIL PROTECTED]> wrote:
I am getting the following messages after a reboot of Solaris 10 FCS
machine:
ufs_log:WARNING: Error writing ufs log state
ufs_log:WARNING: ufs log /net/mercury/u01/backups/log changed state to
Error
ufs_log:WARNING: Please umount(1M) /net/mercury/u01/backups/log and run
fsck(1M)
What do the above messages mean? and how do I get rid off them?
the message itself will appear if the UFS logging code gets I/O
errors of all sorts while reading/writing the ondisk log. This
is considered fatal since we can no longer guarantee transactional
semantics and correctness and thus turn the file system into R/O mode.
Is ufs logging trying to replay some transactions?
/net/mercury/... is a NFS directory so how does ufs logging work for NFS?
this is indeed surprising and without more information I can
only speculate that somehow an attempt was being made to mount
this via UFS or silently something changed underneath of UFS.
what does mount(1M), /etc/vfstab and /etc/mnttan tell you about
the mounted file systems and in particular about this one, is it
really an NFS mounted file system ? also look into the inkernel
state via: `# echo "::fsinfo" |mdb -k`
---
frankB
_______________________________________________
ufs-discuss mailing list
[email protected]
