[LU-3730] sanity-hsm test_3 Error: 'could not create file'

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.5.0
• Labels: HSM revzfs
• Environment: autotest
• Severity: 3
• Assignee: Jinshan Xiong (Inactive)
• Reporter: Keith Mannthey (Inactive)

Description

Sanity-hsm test_3 is failing alot:

One example can be seen here: https://maloo.whamcloud.com/test_sets/0c378cde-ff90-11e2-a3fb-52540035b04c

test_3 	

    Error: 'could not create file'
    Failure Rate: 27.00% of last 100 executions [all branches] 

The only thing of interest I found so far is from the MDS console.
1454 error messages:

05:39:27:Lustre: DEBUG MARKER: == sanity-hsm test 3: Check file dirtyness when opening for write == 05:39:24 (1375879164)
05:39:27:LustreError: 1146:0:(mdt_xattr.c:134:mdt_getxattr_one()) getxattr failed: -61
05:39:27:LustreError: 1146:0:(mdt_xattr.c:134:mdt_getxattr_one()) Skipped 1453 previous similar messages
05:39:27:Lustre: DEBUG MARKER: /usr/sbin/lctl mark  sanity-hsm test_3: @@@@@@ FAIL: could not create file

Tests after this tend to fail as well.

Issue Links

• duplicates: LU-3713 Failure on test suite sanity test_17k: get_xattr_names: llistxattr failed
• duplicates: LU-3669 (mdt_xattr.c:178:mdt_getxattr()) ASSERTION( lu_object_assert_exists(&info->mti_object->mot_obj) ) failed:
• is related to: LU-3704 sanity-hsm test_21 - test_24 Error: 'could not release file'

Comments

[Nathaniel Clark]

This seems to be fixed now that LU-3669 has been closed.

[Nathaniel Clark]

This should be fixed by LU-3669 patch http://review.whamcloud.com/7208/

[Nathaniel Clark]

This appears to be a duplicate of LU-3713

[Jinshan Xiong (Inactive)]

There was no such information to indicate why reading acl access info failed. From the test case itself, I feel it's not necessary to use `cp -p' so changing it to `cp' would work.

[Bruno Faccini (Inactive)]

Jinshan, as you already documented in LU-3704, test_3 failures are due to ENODATA return during [cp -p /etc/passwd $TESTFILE.append || error "could not create file"] command when trying to set "system.posix_acl_access" XATTR. But why are some XATTRs part of "cp -p" work ? And in any case we may just check that file content after copy is ok in the test ?

[James Nunez (Inactive)]

Another failure at: https://maloo.whamcloud.com/test_sets/aa3d4616-001d-11e3-a856-52540035b04c
This one with ZFS.