* Re: [LTP] OOM messages when running full ltp tests on a Hyper-V VM
2015-07-23 11:41 ` Cristian Valean
@ 2015-07-23 11:47 ` Cyril Hrubis
2015-07-23 12:01 ` Jan Stancek
2015-07-23 12:08 ` Li Wang
2 siblings, 0 replies; 6+ messages in thread
From: Cyril Hrubis @ 2015-07-23 11:47 UTC (permalink / raw)
To: Cristian Valean; +Cc: ltp-list
Hi!
> Otherwise, are there are params that can limitor specify the resources that
> LTP can use?
The process in question here is oom01, which is test that checks that
process gets killed if it allocates too much memory.
So the test process forks a child that continues to eat memory until
it's killed.
Generally there are some testcases whose subprocesses gets killed
because they are testing that the system kills processes that allocate
too much memory, acces invalid address, etc.
--
Cyril Hrubis
chrubis@suse.cz
------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [LTP] OOM messages when running full ltp tests on a Hyper-V VM
2015-07-23 11:41 ` Cristian Valean
2015-07-23 11:47 ` Cyril Hrubis
@ 2015-07-23 12:01 ` Jan Stancek
2015-07-23 12:08 ` Li Wang
2 siblings, 0 replies; 6+ messages in thread
From: Jan Stancek @ 2015-07-23 12:01 UTC (permalink / raw)
To: Cristian Valean; +Cc: ltp-list
----- Original Message -----
> From: "Cristian Valean" <cristian.valean@live.com>
> To: ltp-list@lists.sourceforge.net
> Sent: Thursday, 23 July, 2015 1:41:56 PM
> Subject: Re: [LTP] OOM messages when running full ltp tests on a Hyper-V VM
>
> My apologies for the incomplete reply, it got sent out before all details
> went it.
>
> To continue the initial thread:
>
> I’m trying to run a full ltp test run on Ubuntu 14.04.2 as a Hyper-V VM.
>
> The VM is set with 8 or 12GB and 4 vCPUs, however during the tests execution
> the syslog records over 30 call traces similar to the one below.
>
> Haven’t tried vmware or other platforms, but I’m wondering why LTP processes
> go over the limit, or what would be a reference system resources that would
> accommodate a full run.
>
> Otherwise, are there are params that can limitor specify the resources that
> LTP can use?
>
> Sample call trace:
>
> Jul 22 01:17:50 ubuntu kernel: [ 5757.011711] Out of memory: Kill process
> 5902 (oom01) score 937 or sacrifice child
>
> Jul 22 01:17:50 ubuntu kernel: [ 5757.014099] Killed process 5902 (oom01)
> total-vm:15759784kB, anon-rss:7723620kB, file-rss:40kB
oom0* testcases are designed to trigger OOM and verify that expected process
has been killed. These tests go over limit on purpose.
Regards,
Jan
------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [LTP] OOM messages when running full ltp tests on a Hyper-V VM
2015-07-23 11:41 ` Cristian Valean
2015-07-23 11:47 ` Cyril Hrubis
2015-07-23 12:01 ` Jan Stancek
@ 2015-07-23 12:08 ` Li Wang
2 siblings, 0 replies; 6+ messages in thread
From: Li Wang @ 2015-07-23 12:08 UTC (permalink / raw)
To: Cristian Valean; +Cc: ltp-list
Hi,
----- Original Message -----
>
>
> My apologies for the incomplete reply, it got sent out before all details
> went it.
>
>
>
> To continue the initial thread:
>
> I’m trying to run a full ltp test run on Ubuntu 14.04.2 as a Hyper-V VM.
>
> The VM is set with 8 or 12GB and 4 vCPUs, however during the tests execution
> the syslog records over 30 call traces similar to the one below.
It is normal to get these calltrace, you can check the oom0*.c file. These kind of
programs are design to verify the OOM-Killer work well.
$ cat oom01.c
...
*
* The program is designed to cope with unpredictable like amount and
* system physical memory, swap size and other VMM technology like KSM,
* memcg, memory hotplug and so on which may affect the OOM
* behaviours. It simply increase the memory consumption 3G each time
* until all the available memory is consumed and OOM is triggered.
Regards,
Li Wang
>
>
>
> Haven’t tried vmware or other platforms, but I’m wondering why LTP processes
> go over the limit, or what would be a reference system resources that would
> accommodate a full run.
>
> Otherwise, are there are params that can limitor specify the resources that
> LTP can use?
>
>
>
> Sample call trace:
>
> Jul 22 01:17:50 ubuntu kernel: [ 5757.011711] Out of memory: Kill process
> 5902 (oom01) score 937 or sacrifice child
>
> Jul 22 01:17:50 ubuntu kernel: [ 5757.014099] Killed process 5902 (oom01)
> total-vm:15759784kB, anon-rss:7723620kB, file-rss:40kB
>
> Jul 22 01:25:26 ubuntu kernel: [ 6211.981585] min_free_kbytes invoked
> oom-killer: gfp_mask=0x280da, order=0, oom_score_adj=0
>
> Jul 22 01:25:26 ubuntu kernel: [ 6211.982662] min_free_kbytes cpuset=/
> mems_allowed=0
>
> Jul 22 01:25:26 ubuntu kernel: [ 6211.983723] CPU: 1 PID: 12532 Comm:
> min_free_kbytes Not tainted 3.16.0-43-generic #58~14.04.1-Ubuntu
>
>
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.531764] Call Trace:
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.534570] [<ffffffff81765a91>]
> dump_stack+0x45/0x56
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.537888] [<ffffffff8175f645>]
> dump_header+0x7f/0x1f1
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.540655] [<ffffffff811655c5>]
> oom_kill_process+0x205/0x360
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.543361] [<ffffffff812ebbd5>] ?
> security_capable_noaudit+0x15/0x20
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.554536] [<ffffffff81165d72>]
> out_of_memory+0x492/0x4d0
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.557160] [<ffffffff8176019d>]
> __alloc_pages_slowpath+0x63e/0x775
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.559807] [<ffffffff8116bb23>]
> __alloc_pages_nodemask+0x273/0x2a0
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.563601] [<ffffffff811ae93a>]
> alloc_pages_vma+0x9a/0x150
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.569377] [<ffffffff8118f49c>]
> handle_mm_fault+0xcfc/0x10c0
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.574630] [<ffffffff81769c5f>] ?
> __schedule+0x35f/0x7a0
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.582479] [<ffffffff8105af5e>]
> __do_page_fault+0x19e/0x560
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.590664] [<ffffffff810a62e8>] ?
> __enqueue_entity+0x78/0x80
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.597382] [<ffffffff810a8015>] ?
> set_next_entity+0x95/0xb0
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.608215] [<ffffffff81011627>] ?
> __switch_to+0x167/0x580
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.614156] [<ffffffff8105b351>]
> do_page_fault+0x31/0x70
>
> Jul 22 01:17:50 ubuntu kernel: [ 5756.616875] [<ffffffff81770168>]
> page_fault+0x28/0x30
>
>
>
>
> From: Cristian Valean [mailto:cristian.valean@live.com]
> Sent: Thursday, July 23, 2015 2:34 PM
> To: ltp-list@lists.sourceforge.net
> Subject: [LTP] OOM messages when running full ltp tests on a Hyper-V VM
>
>
>
>
> I’m trying to run a full ltp test run on Ubuntu 14.04.2 as a Hyper-V VM.
>
>
>
> The VM is set with 8 or 12GB and 4 vCPUs, however during the tests execution
> the syslog records over 30 call traces similar to the one below.
>
>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/ltp-list
>
------------------------------------------------------------------------------
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list
^ permalink raw reply [flat|nested] 6+ messages in thread