[LU-13240] sanity test 161d fails with 'create should be blocked'

• Type: Bug
• Resolution: Unresolved
• Priority: Minor
• Affects Version/s: Lustre 2.14.0
• Labels: ppc
• Environment: PPC clients
• Severity: 3
• Assignee: WC Triage
• Reporter: James Nunez (Inactive)

Description

sanity test_161d fails with 'create should be blocked' starting on 27 SEPT 2019 and fails 100% of the time for PPC client testing.

Looking at a recent failure at https://testing.whamcloud.com/test_sets/d3720002-4a27-11ea-b69a-52540065bddc, the suite_log shows

== sanity test 161d: create with concurrent .lustre/fid access ======================================= 01:25:08 (1581125108)
CMD: trevis-10vm12 /usr/sbin/lctl get_param mdd.lustre-MDT0000.changelog_mask -n
CMD: trevis-10vm12 /usr/sbin/lctl set_param mdd.lustre-MDT0000.changelog_mask=+hsm
mdd.lustre-MDT0000.changelog_mask=+hsm
CMD: trevis-10vm12 /usr/sbin/lctl --device lustre-MDT0000 changelog_register -n
Registered 1 changelog users: 'cl4'
fail_loc=0x8000140c
fail_val=5
  PID TTY          TIME CMD
 sanity test_161d: @@@@@@ FAIL: create should be blocked 
  Trace dump:
  = /usr/lib64/lustre/tests/test-framework.sh:6121:error()
  = /usr/lib64/lustre/tests/sanity.sh:14593:test_161d()

There is no errors or any indication of why this test fails in any of the console logs.

The code that is failing is

14572         #define OBD_FAIL_LLITE_CREATE_NODE_PAUSE 0x140c | OBD_FAIL_ONCE
14573         $LCTL set_param fail_loc=0x8000140c
14574         # 5s pause
14575         $LCTL set_param fail_val=5
14576 
14577         # create file
14578         echo foofoo > $DIR/$tdir/$tfile &
14579         pid=$!
14580 
14581         # wait for create to be delayed
14582         sleep 2
14583 
14584         ps -p $pid
14585         [[ $? -eq 0 ]] || error "create should be blocked"

Logs for more sanity test 161d failures are at
https://testing.whamcloud.com/test_sets/5e7bd63a-f7af-11e9-b62b-52540065bddc
https://testing.whamcloud.com/test_sets/6a18091c-233a-11ea-bb75-52540065bddc