[LU-6386] lower transno may overwrite the bigger one in client last_rcvd slot

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.8.0
• Severity: 3
• Assignee: Mikhail Pershin
• Reporter: Mikhail Pershin

Description

The tgt_last_rcvd_update() may update transaction number in client slot with smaller value. That may cause client to see wrong error message:

LustreError: 3662:0:(import.c:985:ptlrpc_connect_interpret()) pfs2dat1-OST000a_UUID went back in time (transno 77492683762 was previously committed, server now claims 77492683761)! See https://bugzilla.lustre.org/show_bug.cgi?id=9646

This is possible on OST with several IO requests from one client and safe but it is better to keep transaction value in last_rcvd monotonically increased to avoid confusion on client side.

Issue Links

• is related to: LU-5319 Support multiple slots per client in last_rcvd file

Comments

[Joseph Gmitter (Inactive)]

Landed for 2.8

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/14113/
Subject: LU-6386 tgt: don't update client data with smaller transno
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 5fa93f6011dc37fba3354137457b80151f46c66f

[Supporto Lustre Jnet2000 (Inactive)]

Please could you give us a link where we can download this patch RPM?

Regards

[Andreas Dilger]

It would also be good to change the error message not to reference bugzilla, but instead an LU ticket.

[Gerrit Updater]

Mike Pershin ([e-mail]) uploaded a new patch: http://review.whamcloud.com/14113
Subject: LU-6386 tgt: don't update client data with smaller transno
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: af8cc22454df96abe45e26d7887fd5bc69cc3daf