[LU-17658] sanity check when ofd assign a new sequence to osp

• Type: Improvement
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.17.0
• Severity: 3
• Assignee: Dongyang Li
• Reporter: Dongyang Li

Description

In the case of ofd loses the commit to update sequence range stored in seq_srv,
ofd could assign the same sequence to different osp, creating corruption in the filesystem.

Issue Links

• is related to: LU-17538 lov_objseq file contains 0x0BD0 contstant in low bytes
• is related to: LU-16692 replay-single: test_70c osp_fid_diff()) ASSERTION( fid_seq(fid1) == fid_seq(fid2) )
• is related to: LU-16720 large-scale test_3a osp_precreate_rollover_new_seq()) ASSERTION( fid_seq(fid) != fid_seq(last_fid) ) failed: fid [0x240000bd0:0x1:0x0], last_fid [0x240000bd0:0x3fff:0x0]
• is related to: LU-11912 reduce number of OST objects created per MDS Sequence

Comments

[Peter Jones]

Merged for 2.17

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/54474/
Subject: LU-17658 fid: check on disk sequence before allocating to osp
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 373b76b345b5eaa21b201315075f01ccfc408189

[Andreas Dilger]

I also noticed that conf-sanity test_155 started taking much longer after patch https://review.whamcloud.com/54191 ("LU-17538 fid: do not use allocation set for ofd") landed on 2025-01-02 (average 400s) compared to before landing (average 50s).

There were two different patches with conf-sanity test_155 and each had significantly different test durations, causing the subtest_trend report to be confused.

 

[Gerrit Updater]

"Li Dongyang <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/c/fs/lustre-release/+/54474
Subject: LU-17658 fid: check on disk sequence before allocating to osp
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 4f7c8306d656e0381ea8098df00821489b05a69b