[LU-3852] sanity-hsm test_251: client26-vm "dd: no space left on device"

• Type: Technical task
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.5.0, Lustre 2.7.0
• Labels: HSM
• Assignee: Bruno Faccini (Inactive)
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: http://maloo.whamcloud.com/test_sets/3c70376c-0fbb-11e3-bb21-52540035b04c.

The sub-test test_251 failed with the following error:

request on sanity-hsm is not @@@@@@

== sanity-hsm test 251: Coordinator request timeout == 00:59:34 (1377676774)
Purging archive
Starting copytool
lhsmtool_posix --hsm-root /tmp/arc --daemon --bandwidth 1 /mnt/lustre
dd: writing `/mnt/lustre2/d0.sanity-hsm/d251/f.sanity-hsm.251': No space left on device
93+0 records in
92+0 records out
96620544 bytes (97 MB) copied, 16.7614 s, 5.8 MB/s
CMD: client-27vm3,client-27vm4,client-27vm5,client-27vm6.lab.whamcloud.com /usr/sbin/lctl dk > /logdir/test_logs/2013-08-27/lustre-reviews-el6-x86_64-review-2_4_1_17733_-70153022088660-162844/sanity-hsm.test_251.debug_log.$(hostname -s).1377676791.log;
dmesg > /logdir/test_logs/2013-08-27/lustre-reviews-el6-x86_64-review-2_4_1_17733_-70153022088660-162844/sanity-hsm.test_251.dmesg.$(hostname -s).1377676791.log
CMD: client-27vm3 /usr/sbin/lctl set_param mdt.lustre-MDT0000.hsm_control=disabled
mdt.lustre-MDT0000.hsm_control=disabled
Info required for matching: sanity-hsm 251

Issue Links

• is related to: LU-6055 sanity-hsm file_creation_failure sends df output to stdout
• is related to: LU-4178 Test failure on test suite sanity-hsm, subtest test_200

Comments

[Gerrit Updater]

James Nunez ([e-mail]) uploaded a new patch: http://review.whamcloud.com/13803
Subject: LU-3852 tests: skip tests with large file when no room
Project: fs/lustre-release
Branch: b2_5
Current Patch Set: 1
Commit: d6b1797ac0e0f9ba89688e1fcb986406b65e3418

[Bruno Faccini (Inactive)]

Patch has landed.

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/12456/
Subject: LU-3852 tests: skip tests with large file when no room
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 3a63fe4aa37ed72dc2f9a2c85c35035c2c2619ba

[Bruno Faccini (Inactive)]

I forgot to indicate that change #12456 also re-enables sanity-hsm/test_[200,221,223b,251] sub-tests.

[Bruno Faccini (Inactive)]

Sorry to have been soooo late on this ...
Patch to strengthen sanity-hsm sub-tests, using large files and easy victims of auto-tests configurations with reduced Lustre FS size, is at http://review.whamcloud.com/12456.

[Bruno Faccini (Inactive)]

Patch #9014 to disable test_251 internally in sanity-hsm has landed. So now, will be able to work on a definitive patch to strengthen/protect test_251 vs too small-sized Lustre FS.

[Jodi Levi (Inactive)]

Test has been retriggered: unclear what happened to results

[Bruno Faccini (Inactive)]

Requested Gerrit gate-keeper to land patch #9014, next actions will be to re-allow test_251 in autotest config (thru a new TEI?), and then to push a patch that find a way to handle this "small fs size vs file big enough for timing need" requirement, and also to re-allow test_251 (and also test_[200,221,223b] disabled for the same issue in LU-4178) in same patch.

[Bruno Faccini (Inactive)]

Patch to disable test_251 internally in sanity-hsm has been pushed at http://review.whamcloud.com/9014.
When it will land, I suppose I will need to open a TEI to request current auto-tests config settings to be reversed, like for similar situation with LU-3939/TEI-570/TEI-1041 ?

[Andreas Dilger]

the sanity-hsm test_251 may only be disabled for tests on master, which is ok I think.

Could you please submit a patch to add it to ALWAYS_EXCEPT in the script. When that lands we can remove it from the autotest config so that it will be possible to re-enable it for any patch that is trying to fix the problem.