Bug#516374: task cron blocked for more than 120 seconds.
We have too big problems with this scheduler bug.
It occurs on several Lenny Machines running as VM on ESX3.5U4
Installed Kernel Package is 2.6.26-19
The Machines are not crashing, but they tend to consume lots of cpu
resources until they are nearly unusable or affect other VMs running on
the ESX Host.
This is a "normal" log message from one Vm running a lot of cron jobs:
At the time of writing two cron processes each take 49,9% of cpu :(
regards,
Matthias
[1391701.476841] INFO: task cron:15861 blocked for more than 120 seconds.
[1391701.477846] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs"
disables this message.
[1391701.479429] cron D 464a1a72 0 15861 2938
[1391701.479432] f2167a80 00000286 00000000 464a1a72 000434f2
f2167c0c c180bfa0 00000000
[1391701.479436] 00000000 c01dd0aa 0000084f 00000000 00002400
c01719a0 c011494a c0136114
[1391701.479440] 7fffffff 7fffffff d1d13f68 00000002 c02b89d1
c013458a 00000000 00000000
[1391701.479443] Call Trace:
[1391701.479463] [<c01dd0aa>] __next_cpu+0x12/0x21
[1391701.479520] [<c01719a0>] kmem_cache_alloc+0x50/0x84
[1391701.479525] [<c011494a>] read_real_cycles+0x8/0x9
[1391701.479529] [<c0136114>] getnstimeofday+0x37/0xbc
[1391701.479538] [<c02b89d1>] schedule_timeout+0x13/0x86
[1391701.479542] [<c013458a>] ktime_get_ts+0x22/0x49
[1391701.479547] [<c01345be>] ktime_get+0xd/0x21
[1391701.479552] [<c0119160>] hrtick_start_fair+0xeb/0x12c
[1391701.479561] [<c02b80f5>] wait_for_common+0xaf/0x10f
[1391701.479566] [<c011b6fc>] default_wake_function+0x0/0x8
[1391701.479572] [<c0121c0d>] do_fork+0x17f/0x1dc
[1391701.479581] [<c0102173>] sys_vfork+0x18/0x1c
[1391701.479587] [<c01038ce>] syscall_call+0x7/0xb
[1391701.479599] =======================
Reply to: