[Lxc-users] High SLAB

[China]

Hi,
I've a new issue now with OOM.
I've disabled it for all the containers modifying
/cgroup/<container>/memory.oom_control, but some minutes ago OOM-Killer has
been executed, writing in syslog:

kernel: Out of memory: Kill process 19981 (httpd) score 12 or sacrifice
child
kernel: Killed process 20859 (httpd) total-vm:1022216kB, anon-rss:416736kB,
file-rss:124kB
kernel: httpd invoked oom-killer: gfp_mask=0x0, order=0, oom_adj=0,
oom_score_adj=0
kernel: httpd cpuset=<container> mems_allowed=0
kernel: Pid: 19987, comm: httpd Not tainted 3.3.6 #4
kernel: Call Trace:
kernel: [<ffffffff8110f07b>] dump_header+0x8b/0x1e0
kernel: [<ffffffff8110eb4f>] ? find_lock_task_mm+0x2f/0x80
kernel: [<ffffffff811f93c5>] ? security_capable_noaudit+0x15/0x20
kernel: [<ffffffff8110f8b5>] oom_kill_process+0x85/0x170
kernel: [<ffffffff8110fa9f>] out_of_memory+0xff/0x210
kernel: [<ffffffff8110fc75>] pagefault_out_of_memory+0xc5/0x110
kernel: [<ffffffff81041cfc>] mm_fault_error+0xbc/0x1b0
kernel: [<ffffffff81506873>] do_page_fault+0x3c3/0x460
kernel: [<ffffffff81171253>] ? sys_newfstat+0x33/0x40
kernel: [<ffffffff81503075>] page_fault+0x25/0x30

and killing a process outside the container that has invoked it. But
I've disable OOM-killer completely for containers processes!
I think this is a bug in kernel code.

On Fri, Aug 24, 2012 at 2:23 PM, China <china at email.it> wrote:

> Hi,
> I had misunderstood the proc_sys_vm_vfs_cache_pressure functioning.
> The mechanism works great when the containers need more memory than free
> one. Until that the kernel doesn't free the slab objects.
> Good!
>
> On Fri, Aug 24, 2012 at 1:52 PM, China <china at email.it> wrote:
>
>> Sorry,
>> i've tried that some days ago (setting first to 1000, next 10000 and
>> finally to the maximum1410065408) but the issue remains!
>> My server has about 1800000 files and 500000 directories.
>>
>> Thank you
>>
>>
>> On Fri, Aug 24, 2012 at 1:44 PM, Fajar A. Nugraha <list at fajar.net> wrote:
>>
>>> On Fri, Aug 24, 2012 at 6:37 PM, China <china at email.it> wrote:
>>> > Often, after some hours, the containers occupy a lot of SLAB cache
>>> (over
>>> > 100~200MB), specially dentry and ext3_inode_cache.
>>>
>>> Try this:
>>> http://www.linuxinsight.com/proc_sys_vm_vfs_cache_pressure.html
>>>
>>> I set mine at 1000 for a system with lots of files (> 1M) and
>>> relatively small memory (2GB). Not an lxc system though.
>>>
>>> --
>>> Fajar
>>>
>>
>>
>>
>> --
>>
>> Davide Belloni
>>
>
>
>
> --
>
> Davide Belloni
>



-- 

Davide Belloni
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20120827/3c1e6dd0/attachment.html>