[LU-6725] replay-single test_0a: FAIL: import is not in FULL state

• Type: Bug
• Resolution: Duplicate
• Priority: Blocker
• Affects Version/s: Lustre 2.8.0
• Environment: Lustre Build: https://build.hpdd.intel.com/job/lustre-master/3071; Distro/Arch: RHEL7.1/x86_64
• Severity: 3
• Assignee: Bob Glossman (Inactive)
• Reporter: Jian Yu

Description

replay-single test 0a failed as follows:

Started lustre:MDT0000
CMD: onyx-42vm5,onyx-42vm6.onyx.hpdd.intel.com PATH=/usr/lib64/lustre/tests:/usr/lib/lustre/tests:/usr/lib64/lustre/tests:/opt/iozone/bin:/opt/iozone/bin:/usr/lib64/lustre/tests/mpi:/usr/lib64/lustre/tests/racer:/usr/lib64/lustre/../lustre-iokit/sgpdd-survey:/usr/lib64/lustre/tests:/usr/lib64/lustre/utils/gss:/usr/lib64/lustre/utils:/usr/lib64/openmpi/bin:/usr/bin:/bin:/usr/sbin:/sbin::/sbin:/bin:/usr/sbin: NAME=autotest_config sh rpc.sh wait_import_state_mount FULL mdc.lustre:MDT0000-mdc-*.mds_server_uuid 
onyx-42vm5: CMD: onyx-42vm5.onyx.hpdd.intel.com lctl get_param -n at_max
onyx-42vm6: CMD: onyx-42vm6.onyx.hpdd.intel.com lctl get_param -n at_max
onyx-42vm6:  rpc : @@@@@@ FAIL: can't put import for mdc.lustre:MDT0000-mdc-*.mds_server_uuid into FULL state after 1475 sec, have  
onyx-42vm5:  rpc : @@@@@@ FAIL: can't put import for mdc.lustre:MDT0000-mdc-*.mds_server_uuid into FULL state after 1475 sec, have  

Maloo report: https://testing.hpdd.intel.com/test_sets/d2bcff78-135d-11e5-b4b0-5254006e85c2

Issue Links

• duplicates: LU-6992 recovery-random-scale test_fail_client_mds: test_fail_client_mds returned 254
• is related to: LU-4241 Test failure on test recovery-small test_101: import is not in FULL state

Comments

[Jian Yu]

The label on the MDT device was lustre:MDT0000, which had not been changed to lustre-MDT0000. So, this is a duplicate of LU-6992.

[Saurabh Tandan (Inactive)]

Another instance on 2.7.61 tag:
https://testing.hpdd.intel.com/test_sets/9f06c600-6dee-11e5-b960-5254006e85c2

[Oleg Drokin]

It looks like we see on MDS taht teh recoevry is done so the clients did reconnect.
could it be just proc file update glitch of some sort?

There's no reconnect message from clients to MDS, but there is one for MGS so perhaps its' teh message reduction logic that ate them?
Should be visible in the debug logs then.