[LU-7486] Inconsistent state information in health_check

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.8.0
• Fix Version/s: Lustre 2.9.0
• Type: Bug
• Priority: Minor
• Reporter: Frank Heckes (Inactive)
• Assignee: Bruno Faccini (Inactive)
• Resolution: Fixed
• Labels: soak

Environment

lola
build:2.7.63-4-gf84e06e, a7eface85ea2d2aa6198681264b082a0244855d4

• Severity: 3

Description

The error occurred during soak testing of build '20151122' (see https://wiki.hpdd.intel.com/pages/viewpage.action?title=Soak+Testing+on+Lola&spaceKey=Releases#SoakTestingonLola-20151122).

If a Lustre client hit a LBUG (for example LU-7422) the state information found in /proc/fs/lustre/health_check is inconsistent and confusing:

[root@lola-29 ~]# cat /proc/fs/lustre/health_check
LBUG
healthy

It would be desired to remove the string 'health' or replace it with 'insane' or an equivalent string, mostly to provide meaningful messages to monitoring tools.

Comments

[Bruno Faccini (Inactive)]

Hello Frank, you are right and I just found this too when working with Gabriele to determine health_check proc-file capability!
Will post a patch to report "NOT HEALTHY" like already for any other failure being detected during check.

[Gerrit Updater]

Faccini Bruno ([e-mail]) uploaded a new patch: http://review.whamcloud.com/17981
Subject: LU-7486 obdclass: health_check to report unhealthy upon LBUG
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 45a0dabd513a4c7ab16042a4519aea42263ccc30

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/17981/
Subject: LU-7486 obdclass: health_check to report unhealthy upon LBUG
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 909e4dc00f224834ff7ac4b6b8f0f6bf76e3c58d

[Joseph Gmitter (Inactive)]

patch has landed to master for 2.9.0