[LU-2471] conf-sanity test_24a: mkfs.lustre FATAL: unable to prepare backend

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.4.0
• Affects Version/s: Lustre 2.4.0
• Labels: HB
• Environment: lustre master build #1104 with ZFS
• Severity: 3
• Assignee: Li Wei (Inactive)
• Reporter: Maloo

Description

This issue was created by maloo for sarah <[e-mail]>

This issue relates to the following test suite run: https://maloo.whamcloud.com/test_sets/70b27b00-4295-11e2-8dba-52540035b04c.

The sub-test test_24a failed with the following error:

test_24a returned 10

CMD: client-29vm3 mkfs.lustre --mgsnode=client-29vm3@tcp --fsname=lustre --mdt --index=0 --param=sys.timeout=20 --param=lov.stripesize=1048576 --param=lov.stripecount=0 --param=mdt.identity_upcall=/usr/sbin/l_getidentity --backfstype=ldiskfs --device-size=2097152 --mkfsoptions=\"-E lazy_itable_init\" --mountfsoptions=errors=remount-ro,iopen_nopriv,user_xattr,acl --nomgs --mgsnode=client-29vm3@tcp --fsname=test1234 --reformat lustre-mdt1_2/mdt1_2 /dev/lvm-MDS/S1
client-29vm3: mkfs.lustre: failed to resolve path to lustre-mdt1_2/mdt1_2
client-29vm3:
client-29vm3: mkfs.lustre FATAL: unable to prepare backend (2)
client-29vm3: mkfs.lustre: exiting with 2 (No such file or directory)

Issue Links

• is related to: LU-9043 conf-sanity test cases 24a remove from ALWAYS_EXCEPT

Comments

[Li Wei (Inactive)]

The patch has landed to master. The Autotest issue has been reported to Chris and, I believe, has been fixed in Autotest configurations.

[Li Wei (Inactive)]

http://review.whamcloud.com/4849

[Li Wei (Inactive)]

The format command Sarah quoted is indeed inconsistent:

... --backfstype=ldiskfs ... lustre-mdt1_2/mdt1_2 /dev/lvm-MDS/S1

According the "lustre-initialisation" log in the Maloo session, Autotest gave these:

06:51:51:export MDSFSTYPE=ldiskfs
06:51:51:export MGSFSTYPE=zfs
...
06:51:51:export OSTFSTYPE=zfs
06:51:52:export FSTYPE=zfs

Thus, "--backfstype=ldiskfs" was correct, while mdsdevname() and mdsvdevname() generated incorrect device names. There are two issues need to be fixed here:

1. Autotest should set MDSFSTYPE to "zfs". Typo?
2. facet_type() should recognise inputs like "mds1_2".

I'll submit patch for issue 2.