[LU-9925] mount.lustre: /dev/sda has not been formatted with mkfs.lustre or the backend filesystem type is not supported by this tool

• Type: Bug
• Resolution: Not a Bug
• Priority: Critical
• Affects Version/s: Lustre 2.10.0
• Environment: Lustre: Build Version: 2.10.0_62_ge1d3a0e
• Severity: 3
• Assignee: WC Triage
• Reporter: Brian Murrell (Inactive)

Description

We are occasionally seeing errors trying to mount a Lustre target such as:

# mount -t lustre /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15 /mnt/testfs-OST0000
mount.lustre: /dev/sda has not been formatted with mkfs.lustre or the backend filesystem type is not supported by this tool
# echo $?
19

 

This is after previously having formatted and "registered" (i.e. initial mount) the target:

# mkfs.lustre --ost --mgsnode=10.14.82.168@tcp0 --mgsnode=10.14.82.169@tcp0 --failnode=10.14.82.171@tcp0 --index=0 --mkfsoptions=-J size=2048 --backfstype=ldiskfs --fsname=testfs /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15
# mount -t lustre /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15 /mnt/testfs-OST0000
# umount /dev/sda

 

The kernel log contains after the above:

Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: LDISKFS-fs (sda): file extents enabled, maximum tree depth=5
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: LDISKFS-fs (sda): mounted filesystem with ordered data mode. Opts: errors=remount-ro
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: LDISKFS-fs (sda): file extents enabled, maximum tree depth=5
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: LDISKFS-fs (sda): mounted filesystem with ordered data mode. Opts: ,errors=remount-ro,no_mbcache,nodelalloc
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: Lustre: testfs-OST0000: new disk, initializing
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: Lustre: srv-testfs-OST0000: No data found on store. Initialize space
Aug 29 06:01:44 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: Lustre: testfs-OST0000: Imperative Recovery not enabled, recovery window 300-900
Aug 29 06:01:45 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: Lustre: Failing over testfs-OST0000
Aug 29 06:01:46 lotus-44vm17.lotus.hpdd.lab.intel.com kernel: Lustre: server umount testfs-OST0000 complete

So it sure seems as if the format and registration are successful.

 

However (occasionally, as noted above) when we subsequently [re-]mount that target with the command:

# mount -t lustre /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15 /mnt/testfs-OST0000

we get an error (also as noted at the top of this report):

mount.lustre: /dev/sda has not been formatted with mkfs.lustre or the backend filesystem type is not supported by this tool
# echo $?
19

 

We are certain we have the correct device(s) on this node:

# ls -l /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15
lrwxrwxrwx 1 root root 9 Aug 29 08:50 /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_disk15 -> ../../sda