[Lustre-discuss] Problem mounting file system after upgrading from 1.6.6 -> 1.8.1

Arcand, William warcand at ll.mit.edu
Tue Sep 15 10:35:47 PDT 2009


I have a lustre file system that I recently upgraded from 1.6.6 to 1.8.1
(following the instructions from the Lustre 1.8 manual). After unmounting
the clients, osts and mds I upgraded the kernels/lustre packages. When I
tired to mount the msd I got an error so I ran:

tunefs.lustre --mgs --writeconf --mgs --mdt --fsname=lustrex0 /dev/sda4

This cleared up the error and the mds mounted cleanly. However when I try to
run tune2fs.lustre on my ost, I get the following error:

===
# tunefs.lustre --ost --fsname=lustrex0 --mgsnode=10.23.2.1 at tcp /dev/sda4
checking for existing Lustre data: found CONFIGS/mountdata
Reading CONFIGS/mountdata

   Read previous values:
Target:     lustrex0-OST0002
Index:      2
Lustre FS:  lustrex0
Mount type: ldiskfs
Flags:      0x2
              (OST )
Persistent mount opts: errors=remount-ro,extents,mballoc
Parameters: mgsnode=10.23.2.1 at tcp

tunefs.lustre: cannot change the name of a registered target
tunefs.lustre: exiting with 1 (Operation not permitted)
===


All the previous values shown appear to be correct. If I try to add a
--index=2 I get the same error. I am concerned that my metadata server has
lost it configs, or some other error may be happening.

Any suggestions?

William




More information about the lustre-discuss mailing list