[LU-15151] conf-sanity test_119: mds1: ssh: Could not resolve hostname mds1: Name or service not known

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.15.0
• Severity: 3
• Assignee: Elena Gryaznova
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/4f01fe03-3433-4b2b-9ca9-ebce8607c27b

test_119 PASSED, but

mds1: ssh: Could not resolve hostname mds1: Name or service not known

VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
conf-sanity test_119 -

Issue Links

• is related to: LU-9699 osp_obd_connect()) ASSERTION( osp->opd_connects == 1 ) failed

Comments

[Andreas Dilger]

Alex, what do you have in your config for $mds1_HOST (maps to $mds_HOST if unset)? Can the client $PDSH to that node (should be a no-op if it is a single client).

According to the Gerrit Janitor test logs, the test is taking 1200s to finish, and the wait_update_facet_cond() is continually timing out, but this does not cause an error return:

https://testing-archive.whamcloud.com/gerrit-janitor/19360/results.html

I'm not sure if that is how this test is supposed to work or not, but 1200s is a long time and probably the test should be added to the SLOW list.

[Alex Zhuravlev]

conf-sanity/119 times out every time with the patch:

COMMIT          TESTED  PASSED  FAILED          COMMIT DESCRIPTION
dc27b1c3ea      3       0       3       BAD     LU-15151 tests: use facet check instead of node check
fa36c6b0b9      3       3       0       GOOD    LU-15160 kernel: kernel update SLES12 SP5 [4.12.14-122.91.2]

[Peter Jones]

Landed for 2.15

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/45369/
Subject: LU-15151 tests: use facet check instead of node check
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: dc27b1c3ea1852617012753a3cce1f6c76b164af

[Gerrit Updater]

"Elena Gryaznova <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/45369
Subject: LU-15151 tests: use facet check intead of node check
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 59432387b259d06b0bee1b2e769ee2b639b9c46d

[Andreas Dilger]

It looks like this was introduced by patch https://review.whamcloud.com/27753 "LU-9699 osp: don't assert on OSP duplicating".

The wait_update_cond() call in test_119 should take a hostname as an argument instead of a facet name.