[LU-1570] llog_cat.c:428:llog_cat_process_flags() catlog 0x27500007 crosses index zero

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.1.1
• Type: Bug
• Priority: Minor
• Reporter: Ned Bass
• Assignee: Bob Glossman (Inactive)
• Resolution: Fixed
• Labels: llnl, server

Environment

https://github.com/chaos/lustre/commits/2.1.1-15chaos

• Severity: 3

Description

This message shows up periodically on the MDS of our test filesystem. What needs to be done about it?

Issue Links

• is related to: LU-1586 no free catalog slots for log

Comments

[Bob Glossman (Inactive)]

First look at the code suggests this is just a warning, not a serious error. Can we get a syslog from the MDS? We would like to see if this message is associated with any other behavior we can identify as bad.

[Ned Bass]

There's other "bad" stuff going on in the logs due to recovery testing I was performing. The watchdog timers for mount.lustre and llog_process_th probably need a separate issue opened.

[Ned Bass]

I wonder if this is related? On the same filesystem, we started getting 'Bad address' errors from the client:

On the client:

# hype318 /p/lcrater2/bass6 > dd if=/dev/zero of=foo count=1
dd: opening `baz': Bad address

On the MDS:

Jun 28 12:08:45 zwicky-mds2 kernel: LustreError: 11841:0:(llog_cat.c:81:llog_cat_new_log()) no free catalog slots for log...
Jun 28 12:08:45 zwicky-mds2 kernel: LustreError: 11841:0:(llog_cat.c:81:llog_cat_new_log()) Skipped 3 previous similar messages
Jun 28 12:08:45 zwicky-mds2 kernel: LustreError: 11841:0:(llog_obd.c:454:llog_obd_origin_add()) write one catalog record failed: -28
Jun 28 12:08:45 zwicky-mds2 kernel: LustreError: 11841:0:(llog_obd.c:454:llog_obd_origin_add()) Skipped 3 previous similar messages
Jun 28 12:08:45 zwicky-mds2 kernel: LustreError: 11841:0:(mdd_object.c:1330:mdd_changelog_data_store()) changelog failed: rc=-28 op17 t[0x200de60af:0x17913:0x0]

[Andreas Dilger]

Linking to LU-1586, since this is where the "no free catalog slots for log" message is being discussed.

[Andreas Dilger]

I believe this was fixed with LU-7934.