Attempting to run IOR performance test, one client immediately LBUGS:
[81191.988331] LustreError: 38708:0:(lcommon_cl.c:144:cl_file_inode_init()) ASSERTION( (((inode->i_mode) & 00170000) == 0100000) ) failed:
[81192.032178] LustreError: 38708:0:(lcommon_cl.c:144:cl_file_inode_init()) LBUG
[81192.059027] Pid: 38708, comm: mdtest
[81192.073424]
Call Trace:
[81192.091325] [<ffffffffc092a7ae>] libcfs_call_trace+0x4e/0x60 [libcfs]
[81192.115180] [<ffffffffc092a83c>] lbug_with_loc+0x4c/0xb0 [libcfs]
[81192.137740] [<ffffffffc0e33008>] cl_file_inode_init+0x248/0x300 [lustre]
[81192.162137] [<ffffffffc0e0acf5>] ll_update_inode+0x315/0x600 [lustre]
[81192.185533] [<ffffffff8121dbb9>] ? iget5_locked+0x99/0x210
[81192.205669] [<ffffffffc0d01b75>] ? mdc_get_lustre_md+0x825/0xd80 [mdc]
[81192.229177] [<ffffffffc0e189f0>] ? ll_set_inode+0x0/0x190 [lustre]
[81192.251419] [<ffffffffc0e196c3>] ll_iget+0x163/0x350 [lustre]
[81192.272145] [<ffffffffc0e0d062>] ll_prep_inode+0x232/0xc80 [lustre]
[81192.294489] [<ffffffffc0e1c3b3>] ll_atomic_open+0x783/0x12d0 [lustre]
[81192.317327] [<ffffffff8121006d>] do_last+0xa4d/0x12c0
[81192.335495] [<ffffffff812109a2>] path_openat+0xc2/0x490
[81192.354257] [<ffffffffc0a386c1>] ? lprocfs_counter_sub+0xc1/0x130 [obdclass]
[81192.378986] [<ffffffff81212f3b>] do_filp_open+0x4b/0xb0
[81192.397685] [<ffffffff8122019a>] ? __alloc_fd+0x8a/0x130
[81192.416641] [<ffffffff811ffb83>] do_sys_open+0xf3/0x1f0
[81192.435287] [<ffffffff811ffc9e>] SyS_open+0x1e/0x20
[81192.452763] [<ffffffff811ffcd6>] SyS_creat+0x16/0x20
[81192.470499] [<ffffffff816b5089>] system_call_fastpath+0x16/0x1b
[81192.491356]
[81192.497305] Kernel panic - not syncing: LBUG
Crash dump is available on spirit. This appear identical to some earlier bugs.
any cores left from this would be very long gone. This happened on the old soak test platform. No idea if that hardware even exists now.