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

Re: [Xen-users] dom0 memory usage >1GB at startup (no domU's configured)



I think there is probably more time wasted through investigation of apparent memory issues in OSes than any other performance issue. Frequently these are non-issues.

Here's what you should do:
1. Dedicate 512MB ram and a single VCPU to dom0
2. Restart and continue happily

Sent from my iPhone

On Apr 4, 2010, at 7:52 AM, stalks <xen-users@xxxxxxxxxxxxxxxxxx> wrote:

On 04/04/2010 03:44, stalks wrote:
Dear list,

I have a new install of Debian Squeeze.

I have followed the hints from the ParaVirtOps wiki page to grab Jeremy's kernel sources from kernel.org and compiled a 2.6.31.x (and 2.6.33.x) kernel.

I have yet to install any significant software other than package "xen-hypervisor-3.4-amd64".

The machine is a Core i7 920 with 8GB RAM.

If I boot the plain kernel (without xen-3.4.gz) I get ~65MB ram usage shown in free -m.

If I boot with full xen, bear in mind I have no domU's yet, I get the following:

> free -m
total used free shared buffers cached Mem: 7948 1412 6535 0 9 107
-/+ buffers/cache:       1295       6652
Swap:            0          0          0

> xm info
host                   : oracle
release                : 2.6.33.2-amd64-xen
version                : #1 SMP Sun Apr 4 00:21:54 BST 2010
machine                : x86_64
nr_cpus                : 8
nr_nodes               : 1
cores_per_socket       : 4
threads_per_core       : 2
cpu_mhz                : 2660
hw_caps : bfebfbff: 28100800:00000000:00000340:0098e3bd:00000000:00000001:00000000
virt_caps              : hvm
total_memory           : 8183
free_memory            : 2174
node_to_cpu            : node0:0-7
node_to_memory         : node0:2174
xen_major              : 3
xen_minor              : 4
xen_extra              :
xen_caps : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0- x86_32 hvm-3.0-x86_32p hvm-3.0-x86_64
xen_scheduler          : credit
xen_pagesize           : 4096
platform_params        : virt_start=0xffff800000000000
xen_changeset          : unavailable
cc_compiler            : gcc version 4.4.3 (Debian 4.4.3-2)
cc_compile_by          : waldi
cc_compile_domain      : debian.org
cc_compile_date        : Mon Mar  1 21:16:18 UTC 2010
xend_config_format     : 4

> xm top
  xentop - 03:33:21   Xen 3.4
1 domains: 1 running, 0 blocked, 0 paused, 0 crashed, 0 dying, 0 shutdown Mem: 8379516k total, 6152576k used, 2226940k free CPUs: 8 @ 2660MHz

(free -m): Over 1GB used inside dom0 already?
(xm info): Only 2GB available for guests?
(xm top): Over 6GB used?

The only processes running outside of bare minimum are xinetd, sendmail, mdadm and ssh.

I can stop xend and xendomains from starting, and I still have 1GB used at startup.

The same setup on a lesser spec'd machine is working fine with ~140Mb used at boot. I've spent a few hours on this and am at a loss.

Some info can be found at the following sources:

http://www.nooblet.org/other/xen-users/lspci.txt
http://www.nooblet.org/other/xen-users/lspci-vv.txt
http://www.nooblet.org/other/xen-users/cpuinfo.txt
http://www.nooblet.org/other/xen-users/grub.txt
http://www.nooblet.org/other/xen-users/dmesg.txt
http://www.nooblet.org/other/xen-users/meminfo.txt
http://www.nooblet.org/other/xen-users/ps-awwux.txt

Someone show me the light!

Regards,
Steve.

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


I have just tried 4.0.0 rc9 (xen-unstable.hg) and 2.6.32.10 with the same results.

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


_______________________________________________
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®.