[LU-17854] Router should not drop message past deadline

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.16.0
• Severity: 3
• Assignee: Chris Horn
• Reporter: Chris Horn

Description

Router logs show a lot of messages being dropped because they have exceeded their deadline:

$ grep -c 'exceeded message deadline' dklog.router.*206*.fmt
dklog.router.485.20240206_113601.fmt:1092
dklog.router.486.20240206_113601.fmt:261
dklog.router.93.20240206_113601.fmt:1127
dklog.router.94.20240206_113601.fmt:1107
$

This is new code added by LU-16303. It is possible that this is causing a regression in this case as some traffic that would normally be forwarded successfully by the router is instead being dropped.

This ticket will track removing this behavior for messages being forwarded.

Comments

[Peter Jones]

Merged for 2.16

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/55131/
Subject: LU-17854 lnet: Router should not drop msg past deadline
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 182b0101d3b803173040de2c69e41955fd6728b2

[Gerrit Updater]

"Chris Horn <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/c/fs/lustre-release/+/55131
Subject: LU-17854 tests: Router should not drop msg past deadline
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 85ee2150e7af31d026d859dd294dadbfa8f52161