[LU-12674] osp should handle -EINPROGRESS on llog objects

• Status: Resolved
• Project: Lustre
• Affects Version/s: Upstream
• Fix Version/s: Lustre 2.13.0, Lustre 2.12.4
• Type: Bug
• Priority: Minor
• Reporter: Alex Zhuravlev
• Assignee: Alex Zhuravlev
• Resolution: Fixed
• Severity: 3

Description

osp_sync_declare_add() panics if llog isn't ready yet (e.g. because of OI scrub in the middle of fixing llog objects):

[ 5325.840390] LustreError: 27380:0:(osp_sync.c:283:osp_sync_declare_add()) ASSERTION( ctxt ) failed:
[ 5325.850491] LustreError: 27380:0:(osp_sync.c:283:osp_sync_declare_add()) LBUG
[ 5325.858448] Pid: 27380, comm: tgt_recov
[ 5325.858449]
Call Trace:
[ 5325.858468]  [<ffffffffc09c480e>] libcfs_call_trace+0x4e/0x60 [libcfs]
[ 5325.858473]  [<ffffffffc09c4dcc>] lbug_with_loc+0x4c/0xc0 [libcfs]
[ 5325.858487]  [<ffffffffc139fd74>] osp_sync_declare_add+0x374/0x3b0 [osp]
[ 5325.858493]  [<ffffffffc138cc03>] osp_declare_object_destroy+0x103/0x150 [osp]
[ 5325.858505]  [<ffffffffc12e44c9>] lod_declare_object_destroy+0x309/0x870 [lod]
[ 5325.858519]  [<ffffffffc133b789>] mdd_declare_finish_unlink+0xb9/0x1e0 [mdd]
[ 5325.858525]  [<ffffffffc134274c>] mdd_unlink+0x48c/0xa90 [mdd]
[ 5325.858542]  [<ffffffffc1236556>] mdo_unlink+0x46/0x48 [mdt]
[ 5325.858551]  [<ffffffffc11ffcf5>] mdt_reint_unlink+0xa85/0x11d0 [mdt]
[ 5325.858560]  [<ffffffffc12038a3>] mdt_reint_rec+0x83/0x210 [mdt]
[ 5325.858567]  [<ffffffffc11e3d09>] mdt_reint_internal+0x5d9/0xb30 [mdt]
[ 5325.858575]  [<ffffffffc11eedb7>] mdt_reint+0x67/0x140 [mdt]
[ 5325.858630]  [<ffffffffc0e734f0>] tgt_request_handle+0x900/0x11f0 [ptlrpc]
[ 5325.858672]  [<ffffffffc0dc2c4d>] handle_recovery_req+0x8d/0x270 [ptlrpc]
[ 5325.858689]  [<ffffffffc0dc5d6a>] target_recovery_thread+0x7ba/0x1c50 [ptlrpc]
[ 5325.858710]  [<ffffffff810b4031>] kthread+0xd1/0xe0

Instead OSP should skip new llog record and make some noise in the kernel log.

Issue Links

• is related to: LU-13452 MDT is 100% full, cannot delete files
• is related to: LU-14056 MDT filesystem full, fsck crashing

Comments

[Gerrit Updater]

Alex Zhuravlev ([e-mail]) uploaded a new patch: https://review.whamcloud.com/35844
Subject: LU-12674 osp: handle -EINPROGRESS on llog objects
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 6fdf6f009d5ac64cd4e1b9893d9e70bce2f7518d

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/35844/
Subject: LU-12674 osp: handle -EINPROGRESS on llog objects
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: a3ec8ff69fceb53a467a80c2e6008869f25f72b4

[Peter Jones]

Landed for 2.13

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: https://review.whamcloud.com/36348
Subject: LU-12674 osp: handle -EINPROGRESS on llog objects
Project: fs/lustre-release
Branch: b2_12
Current Patch Set: 1
Commit: 8168cb152a2813dec26850fb5ce570e820921298

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/36348/
Subject: LU-12674 osp: handle -EINPROGRESS on llog objects
Project: fs/lustre-release
Branch: b2_12
Current Patch Set:
Commit: f0cf4dfa339f0252cda73f57fe833012c3aea4fa

[Gerrit Updater]

Alex Zhuravlev ([e-mail]) uploaded a new patch: https://review.whamcloud.com/38224
Subject: LU-12674 osp: handle -EINPROGRESS on llog objects
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: d908f25de1977803d1c734bb029518e6c72a88f9