On Aug 08, 2008 13:27 -0300, Balagopal Pillai wrote:> The storage server that served Lustre had a drive
> failure on the 3ware 9650 raid6 array and it seemed to
> rebuild fine. But almost all volumes in that array seems to have
> filesystem corruption. The partitions with OS
> didn''t boot after i did the fsck as it deleted lots of inodes
(entire
> directory with lustre kernel and lots of config files went missing!).
> Now the server is up and running again and i am trying to recover the
> lustre volumes. After the fsck, the mdt mounted fine, but ost
doesn''t
> mount. It gives the following errors. The running version of Lustre
> before was 1.6.0.1 on centos 4 x86_64. The new install is centos 5
> x86_64 and lustre version 1.6.5.1 -
>
> mount.lustre: mount /dev/sda7 at /ost-home failed: No such file or
directory
> Is the MGS specification correct?
> Is the filesystem name correct?
> If upgrading, is the copied client log valid? (see upgrade docs)
>
>
> (from dmesg)
> LDISKFS FS on sda7, internal journal
> LDISKFS-fs: mounted filesystem with ordered data mode.
> LustreError: 3202:0:(obd_mount.c:247:ldd_parse()) cannot open
> CONFIGS/mountdata: rc = -2
> LustreError: 3202:0:(obd_mount.c:1264:server_kernel_mount()) premount
> parse options failed: rc = -2
> LustreError: 3202:0:(obd_mount.c:1560:server_fill_super()) Unable to
> mount device /dev/sda7: -2
> LustreError: 3202:0:(obd_mount.c:1951:lustre_fill_super()) Unable to
> mount (-2)
>
>
> /dev/sda7 does exist. So i have decided against running fsck on
> the other bigger Lustre volumes for now.
> Is it possible to get sda7 mounted any other way? Thanks very much for
> any suggestions and help.
It appears the missing file is "CONFIGS/mountdata" (possibly among
other
missing files). You can maybe recreate this with a --writeconf, depending
on what the rest of the filesystem looks like.
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.