[LU-10347] sanity-hsm test_252: archive request fails rather than canceling out

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.11.0
• Severity: 3
• Assignee: WC Triage
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/42d77f44-db21-11e7-a066-52540065bddc.

The sub-test test_252 failed with the following error:

request on 0x200000405:0x133:0x0 is not CANCELED on mds1

Info required for matching: sanity-hsm 252

Comments

[Joseph Gmitter (Inactive)]

Fix to suspend has landed for 2.11.0

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/30492/
Subject: LU-10347 tests: suspend the copytool in sanity-hsm/test_252
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 96fbf0935977a9669d2a3bb2612db8b7eba3e5a5

[Gerrit Updater]

Quentin Bouget ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30492
Subject: LU-10347 tests: suspend the copytool in sanity-hsm/test_252
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 6ec2a51bffe5a9d5b7e0b5a8068a95fa2780a369

[Bob Glossman (Inactive)]

another on master:
https://testing.hpdd.intel.com/test_sets/ac23e53a-dbd2-11e7-9840-52540065bddc

[Quentin Bouget (Inactive)]

No, adding a delay is not reliable enough. The copytool just needs to be suspended until the request times out (although in that case you will hit LU-10302).

[Gerrit Updater]

John L. Hammond ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30434
Subject: LU-10347 test: give CT time to open file
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 2f3dda96a4d1e6480b5f909a0820ae5a5f720a13

[John Hammond]

The CT calls ct_begin() before opening the file to be archived so there is a small race in this test:

	$LFS hsm_archive --archive $HSM_ARCHIVE_NUMBER $f
        wait_request_state $fid ARCHIVE STARTED
	rm -f $f

which cause the archive request to be failed rather than canceled.