[LU-5441] Client logs broken after update to 2.5.2

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.5.2
• Type: Bug
• Priority: Major
• Reporter: Oliver Mangold
• Assignee: Peter Jones
• Resolution: Not a Bug
• Severity: 3

Description

I updated a Lustre installation from Lustre 2.1.6 to 2.5.2. Now something seems to be wrong with the llogs. After doing writeconf on all targets I can mount a client once, but after unmounting the client again, all subsequent tries to mount clients result in errors like this:

2014-07-31T12:57:12.198389+02:00 l3mds1 <kern.err> kernel:LustreError: 68983:0:(obd_mount.c:1323:lustre_fill_super()) Unable to mount (-5)
2014-07-31T12:57:13.198412+02:00 l3mds1 <kern.err> kernel:LustreError: 15c-8: MGC10.3.6.21@o2ib: The configuration from log 'lustre3-client' failed (-5). This may be the result of communication errors between this node and the MGS, a bad configuration, or other errors. See the syslog for more information.
2014-07-31T12:57:13.198428+02:00 l3mds1 <kern.err> kernel:LustreError: 69606:0:(llite_lib.c:1046:ll_fill_super()) Unable to process log: -5

Can the client logs be brought back into a consistent state? Is writeconf not supposed to fix any llogs issues?

Comments

[Peter Jones]

Sorry to hear about the troubles with your upgrade Oliver.

[Erich Focht]

It turned out that the llogs were incomplete, only 9 of the 48 OSTs actually had llogs on the MGS. Also the client llog was incomplete. The procedure leading to the trouble was:

• delete llogs with tunefs.lustre --writeconf on MDT and OSTs
• mount MDT
• mount OSTs all at once (more or less, controlled by HA)

Correct results were achieved when the third step was done mounting each OST in order, one by one.

While there is still the question: "why does this happen?", I think we can close this ticket.

[John Fuchs-Chesney (Inactive)]

Thanks Erich!

Best regards,
~ jfc.