WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-users

[Xen-users] understanding dom0 usage memory

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] understanding dom0 usage memory
From: Claudinei Matos <claudineimatos@xxxxxxxxxxxxxxxxx>
Date: Fri, 15 Feb 2008 14:11:09 -0200
Delivery-date: Fri, 15 Feb 2008 08:12:07 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Thunderbird 2.0.0.9 (X11/20071031)
Hi guys,

First of all, sorry for my poor english.

This is my first message at this list. I'm using Xen on several machines since last November. Right now I'm having a specific problems with 2 servers: both are losing the network on dom0 but all guest systems are still running fine.

Looking at those dom0 logs I've discovered that sometimes linux kernel OOM-Killer tries to kill some important applications like sshd for e.g. Using ps, top, pmap and free commands I can see that dom0 will try to eat as much memory as there is available, but I can't see all the process using this memory, even if I do sum the virtual memory size (VSZ) of each process, which is supposed to be as big as the total machine (dom0) memory AFAIK.

For example, I have 4 AMD64 systems running Gentoo Linux with Xen 3.1.1 and linux sources 2.6.20-r6.
Two of those machines have 512MB for dom0 and the others have 430MB.
Those hypervisors are eating around 400MB to 500MB (from `free -m` 'used' output) each one, and the only process I'm running on them are mdadm (for raid monitor) ntpd and sshd. If I sum those machines' processes, I can see between 130 to 200 MB of VSZ, so where's the memory not being used? If I do sum RSS (resident memory) plus SHR (shared memory) for all process I'll get somethiing between 80MB to 120MB being used.

On a test machine with only 128MB available to dom0 running the same processes I can see around 117MB in use. On other machines without Xen I can sum all process RSS plus SHR and I'll get almost the same value from `free -m` output.

My question is if the xen kernel part do try to allocate memory which I can't see on ps/top utils or if I maybe have a problem with my machines setup? Another question is if it's OK to just allocate 128MB to dom0? As I said the only machine with 128MB is a test machine which is not used this much.

Thanks for any advice,

Claudinei Matos

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

<Prev in Thread] Current Thread [Next in Thread>