fsck runs without errors but fs won't mount r/w

[news groups]

The computer took a power hit and when it rebooted one of the filesystems was in this strange state. It passes full fsck, but still will not mount read-write, just read-only. Any suggestions? (Serious replies only, thanks.)
# uname -a
HP-UX nodename B.10.20 E 9000/800 454506301 8-user license
# grep maxdsiz /stand/system
maxdsiz 943718400
# ulimit -d
921600
# mount | grep rep24
/repository/rep24 on /dev/vgfbr2/lvol6 ro on Wed Jun 7 15:29:08 2000
# gdf -h /repository/rep24
Filesystem Size Used Avail Capacity Mounted on
/dev/vgfbr2/lvol6 125G 112G 12G 90%
/repository/rep24
# gdf -i /repository/rep24
Filesystem Inodes IUsed IFree %IUsed Mounted on
/dev/vgfbr2/lvol6 11925116 8527392 3397724 72%
/repository/rep24
# umount /repository/rep24
# fsck -F vxfs -o full /dev/vgfbr2/rlvol6
log replay in progress
pass0 - checking structural files
pass1 - checking inode sanity and blocks
pass2 - checking directory linkage
pass3 - checking reference counts
pass4 - checking resource maps
OK to clear log? (ynq)y
set state to CLEAN? (ynq)y
# echo $?
0
# mount -F vxfs -o rw /dev/vgfbr2/lvol6 /repository/rep24 vxfs mount: /dev/vgfbr2/lvol6 is not a vxfs file system. (presumably EINVAL)
# mount -F vxfs -o ro /dev/vgfbr2/lvol6 /repository/rep24
#

[Anthony Goonetilleke]

This is a bit of a twisted method but if you have enough space, mount it read only copy everything on it to another disk then do a pvcreate on the old disk and recreate the LV and recopy the data back.

[Alan Riggs]

Is is possible that the log area for the filesystem is damaged. Try:

fsck -o nolog

[Paul Hite]

Do you have PHKL_21595 or some other previous patch that addresses SR 1653302711? Without that patch a mount of a vxfs filesystem will fail if the filesystem has more the 8,000,000 files.