[LU-15284] lmv_get_fid_from_lsm()) ASSERTION( lmv_dir_striped(lsm) ) failed

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.16.0
• Affects Version/s: Lustre 2.15.0
• Severity: 3
• Assignee: Lai Siyao
• Reporter: Oleg Drokin

Description

This looks like a relatively recent crash introduced sometime around August 2021, since it was first seen on Aug 18 in master-next

It hits in racer like this, the full stack below lookup_it_finish varies:

LustreError: 30377:0:(lmv_obd.c:3710:lmv_get_fid_from_lsm()) ASSERTION( lmv_dir_striped(lsm) ) failed:
LustreError: 30377:0:(lmv_obd.c:3710:lmv_get_fid_from_lsm()) LBUG
Pid: 30377, comm: ls 3.10.0-7.9-debug #1 SMP Wed Nov 17 15:17:29 EST 2021
Call Trace:
[<0>] libcfs_call_trace+0x90/0xf0 [libcfs]
[<0>] lbug_with_loc+0x4c/0xa0 [libcfs]
[<0>] lmv_get_fid_from_lsm.part.30+0x3b/0x3b [lmv]
[<0>] lmv_get_fid_from_lsm+0x23/0xa0 [lmv]
[<0>] ll_lookup_it_finish.isra.37+0xeff/0xf10 [lustre]
[<0>] ll_lookup_it.constprop.39+0xcb2/0x1200 [lustre]
[<0>] ll_atomic_open+0x1b4/0x1800 [lustre]
[<0>] do_last+0xa40/0x1280
[<0>] path_openat+0xcd/0x5b0
[<0>] do_filp_open+0x4d/0xb0
[<0>] do_sys_open+0x124/0x220
[<0>] SyS_openat+0x14/0x20
[<0>] system_call_fastpath+0x1f/0x24
[<0>] 0xfffffffffffffffe 

Sample crash with vmcore and all: https://testing.whamcloud.com/gerrit-janitor/external/crashes/boilpot-bigmem-65-2021-11-27-11:19:48