[Lustre-discuss] soft lockup detected

Bernd Schubert bs at q-leap.de
Thu Oct 4 10:18:25 PDT 2007


Hi,

on deleting a directory with quite a lot of files we get a soft lockup, but 
the stack trace does look rather strange. Somehow I think this looks more 
like a timer/hardware bug than like a lustre bug, doesn't it? Unfortunately 
I can't test with another clocksource, only jiffies are supported.
Any ideas?


[ 2693.454376] Call Trace:
[ 2693.458538]  [<ffffffff8020b40b>] dump_trace+0xeb/0x471
[ 2693.463971]  [<ffffffff8020b83d>] show_trace+0x49/0x68
[ 2693.469288]  [<ffffffff8020b877>] dump_stack+0x1b/0x1d
[ 2693.474579]  [<ffffffff8025fcb8>] softlockup_tick+0x102/0x118
[ 2693.480563]  [<ffffffff8023a569>] run_local_timers+0x13/0x15
[ 2693.486399]  [<ffffffff8023a5be>] update_process_times+0x53/0x85
[ 2693.492630]  [<ffffffff80216af5>] smp_local_timer_interrupt+0x36/0x56
[ 2693.499306]  [<ffffffff80216b7f>] smp_apic_timer_interrupt+0x6a/0x87
[ 2693.505860]  [<ffffffff8020abab>] apic_timer_interrupt+0x6b/0x70
[ 2693.512120]  [<ffffffff8812f172>] :libcfs:cfs_alloc_flags_to_gfp+0x2/0x40
[ 2693.519210]  [<ffffffff8812f1d1>] :libcfs:cfs_alloc+0x21/0x60
[ 2693.525254]  [<ffffffff88164f47>] :lnet:LNetMEAttach+0xe7/0x300
[ 2693.531437]  [<ffffffff8821e4c3>] :ptlrpc:ptl_send_rpc+0x5d3/0xd90
[ 2693.537909]  [<ffffffff882135bb>] :ptlrpc:ptlrpc_send_new_req+0x3fb/0x540
[ 2693.545017]  [<ffffffff88216674>] :ptlrpc:ptlrpc_check_set+0x134/0xaf0
[ 2693.551897]  [<ffffffff8823b74f>] :ptlrpc:ptlrpcd_check+0x18f/0x2b0
[ 2693.560350]  [<ffffffff8823b9f5>] :ptlrpc:ptlrpcd+0x185/0x4a0
[ 2693.566351]  [<ffffffff8020ad88>] child_rip+0xa/0x12
[ 2693.571515]
[ 2703.572278] BUG: soft lockup detected on CPU#0!
[ 2703.577042]
[ 2703.577043] Call Trace:
[ 2703.581162]  [<ffffffff8020b40b>] dump_trace+0xeb/0x471
[ 2703.586629]  [<ffffffff8020b83d>] show_trace+0x49/0x68
[ 2703.592036]  [<ffffffff8020b877>] dump_stack+0x1b/0x1d
[ 2703.597347]  [<ffffffff8025fcb8>] softlockup_tick+0x102/0x118
[ 2703.603323]  [<ffffffff8023a569>] run_local_timers+0x13/0x15
[ 2703.609240]  [<ffffffff8023a5be>] update_process_times+0x53/0x85
[ 2703.615444]  [<ffffffff80216af5>] smp_local_timer_interrupt+0x36/0x56
[ 2703.622076]  [<ffffffff80216b7f>] smp_apic_timer_interrupt+0x6a/0x87
[ 2703.628685]  [<ffffffff8020abab>] apic_timer_interrupt+0x6b/0x70
[ 2703.634905]  [<ffffffff8036799d>] vsnprintf+0xa1/0x64f
[ 2703.640282]  [<ffffffff8813511d>] :libcfs:libcfs_debug_vmsg2+0x4fd/0x990
[ 2703.647206]  [<ffffffff882131a5>] :ptlrpc:ptlrpc_import_delay_req+0x215/0x230
[ 2703.654491]  [<ffffffff88216963>] :ptlrpc:ptlrpc_check_set+0x423/0xaf0
[ 2703.661177]  [<ffffffff8823b74f>] :ptlrpc:ptlrpcd_check+0x18f/0x2b0
[ 2703.667610]  [<ffffffff8823b9f5>] :ptlrpc:ptlrpcd+0x185/0x4a0
[ 2703.673486]  [<ffffffff8020ad88>] child_rip+0xa/0x12
[ 2703.678578]


This is with lustre-1.4.11.


Thanks,
Bernd

-- 
Bernd Schubert
Q-Leap Networks GmbH




More information about the lustre-discuss mailing list