[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] dom0 got load of 80


  • To: "Fajar A. Nugraha" <fajar@xxxxxxxxx>
  • From: Heiko <rupertt@xxxxxxxxx>
  • Date: Wed, 28 Jan 2009 09:30:33 +0100
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Wed, 28 Jan 2009 00:32:32 -0800
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; b=WwXhFJ+3il33eWHrc3ySPF1DDqHf6fi8BPLJa/o8T28tcTy013r7rIKE3KaiqR9rcD IScOuQYcQ/fy1wR96t8p2TGOopLXbKfLqVLTstyuNKxxB+/2BB+gGPiNLplC2RggR3Kc tDUXuqjTvG1HA9Qwnjpa5S8YuXqhQG8Bjz2m8=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

On Wed, Jan 28, 2009 at 9:15 AM, Fajar A. Nugraha <fajar@xxxxxxxxx> wrote:
> On Wed, Jan 28, 2009 at 2:49 PM, Heiko <rupertt@xxxxxxxxx> wrote:
>> Can it be that one domU has that much load that it can take down all
>> the others inlc. dom0?
>
> It should not. However it depends on how you set it up. Can you share
> how you setup dom0 and domU, in particular CPU and memory allocation?
>
Hello,

i dont have set any option to give dom0 a seperate CPU or RAM,
do I have to set  (dom0-cpus 0) to (dom0-cpus 1) for that?
For the memory i saw a option that goes with the grub command.

these are my domU on that host:

[root@x1blade1:~]# xm list
Name                                      ID Mem(MiB) VCPUs State   Time(s)
Domain-0                                   0    10147     8 r----- 388610.5
auto-input-vm1                             1      999     2 -b----  10847.7
distribution-vm1                           5      999     2 -b---- 145234.6
monitoring-1                               2     1999     2 -b---- 1060094.1
translator-vm1                             4      999     2 -b----  63999.2
uat-vm1                                    3      999     2 -b----   9666.5



> A "good" dom0 setup would tipically :
> - has its own dedicated CPU (usually dom0), not used by domUs.
> - has enough memory (512MB would do)
> - not running any unneeded services (e.g. not running nfs server, http
> server, etc.)
>
>> Jan 28 06:07:32 x1blade1 kernel: python invoked oom-killer:
>> gfp_mask=0x201d2, order=0, oomkilladj=0
>> Jan 28 06:07:32 x1blade1 kernel:
>> Jan 28 06:07:32 x1blade1 kernel: Call Trace:
>> Jan 28 06:07:32 x1blade1 kernel:  [<ffffffff802b4896>] 
>> out_of_memory+0x8b/0x203
>> Jan 28 06:07:32 x1blade1 kernel:  [<ffffffff8020f05e>] 
>> __alloc_pages+0x22b/0x2b4
>
> This is bad. Are you perhaps running lots of services on dom0?
>
just xen and nagios nrpe which checks some system info and the domU
for there status.
in this "top" output you can see that python does something:

top - 08:25:00 up 11 days, 11:51,  1 user,  load average: 0.27, 0.28, 0.37
Tasks: 196 total,   1 running, 195 sleeping,   0 stopped,   0 zombie
Cpu(s):  5.5%us,  7.1%sy,  0.0%ni, 84.7%id,  0.5%wa,  0.0%hi,  1.0%si,  1.3%st
Mem:  10390528k total,   937536k used,  9452992k free,   267344k buffers
Swap:  2048276k total,    10380k used,  2037896k free,    77120k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 6976 root      16   0  8756 1116  424 S   50  0.0 884:31.08 xenstored
  368 root      15   0  368m 7000 1296 S   17  0.1   1:12.43 python
 7329 root      18   0  143m 7316 2656 S    6  0.1   0:00.06 python
 7342 root      18   0  143m 7312 2656 S    5  0.1   0:00.05 python
 7351 root      18   0  143m 7312 2656 S    5  0.1   0:00.05 python


greetings

Heiko
> Regards,
>
> Fajar
>

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.