Hi,
Thanks for answer.
I’ll try your suggestions about changes in configuration.
Don’t know if
amd64 kernel will run stable, as this is dual Intel Xeon machine…
My configuration is
very simple, example of one machine: dom0 is just Xen host with network
configured as custom script using two bridges (eth0, eth1). Root partition has
10GB, rest is managed by LVM for DomU’s. It has ~700MB ram. That’s
all, there are no other services on Dom0
DomU’s are
created to provide services:
DomU 1: MySQL 5.1
(from lenny-backports), ext3 (former I have tried XFS here with same effect),
3GB ram
DomU 2: Web (Apache2,
php5 as fastcgi, vsftpd, dkim-filter for emails sent from scripts), XFS, 2GB
ram
DomU 3: Reverse proxy
(Nginx, ftp_proxy, awstats as fastcgi), XFS, 1.2GB ram
Every domain (dom0,
domU’s) have postfix as MTA
Everything installed
via apt-get, no compilation, just default apt repository + lenny-backports in
some cases… 3rd DomU (reverse proxy) is most stable. It
haven’t crash at all. Others crash at least one of them in week…
If you need any other
information – just let me know.
--
Maciek
From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
[mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Peter Viskup
Sent: Tuesday, July 06, 2010 9:21 PM
To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-users] Xen 3.2.1-2 on Debian Lenny 2.6.26 2.6.26-24
On 07/06/2010 07:06 PM, Maciej Liżewski wrote:
Hi,
Recently I have installed Debian Lenny on
two different machines (different ram size, disks, Xeon dual and quad core,
filesystems both xfs and ext3, etc). Packages versions:
Dom0:
ii
libc6-xen
2.7-18lenny4
GNU C Library: Shared libraries [Xen version]
ii
libxenstore3.0
3.2.1-2
Xenstore communications library for Xen
ii
linux-image-2.6.26-2-xen-686
2.6.26-24
Linux 2.6.26 image on i686, oldstyle Xen support
ii
linux-modules-2.6.26-2-xen-686
2.6.26-24
Linux 2.6.26 modules on i686
ii
xen-hypervisor-3.2-1-i386
3.2.1-2
The Xen Hypervisor on i386
ii xen-linux-system-2.6.26-2-xen-686
2.6.26-24
XEN system with Linux 2.6.26
image on i686
ii
xen-shell
1.8-3
Console based Xen administration utility
ii
xen-tools
3.9-4
Tools to manage Debian XEN virtual servers
ii
xen-utils-3.2-1
3.2.1-2
XEN administrative tools
ii
xen-utils-common
3.2.0-2
XEN administrative tools - common files
ii xenstore-utils
3.2.1-2
Xenstore utilities for Xen
domU’s:
ii
libc6-xen
2.7-18lenny4
GNU C Library: Shared libraries [Xen version
ii
linux-modules-2.6.26-2-xen-686
2.6.26-24 Linux
2.6.26 modules on i686
those are common for both installations.
Anly standard packages were installed using “apt-get”. No
compiling, no unstable releases…. There are 2-3 DomU’s on each
machine. There have very little usage (load average under 1.0). Both
installations was made about month ago. During this time domU’s are
randomly crashing. Interval between those crashes vary between single day and
week. There are no logs about reason of those crashes, just “Domain has
crashed” in xend.log and after that log about xen trying to respawn DomU
but previous instance is still running. Only a single DomU on machine is
crashing at a time – others run stable at the same time. Which DomU will
crash next time is undeterminable and random…
Funny, that times of most those crashes are
similar to MRTG executed from CRON (and log about this execution is often last
in syslog just before crash), however there are nothing specific in MRTG config
(getting cpu usage, load average, disk space) but using piped commands instead
of snmp (they are simple: cat /proc/meminfo | grep | awk… and so on)
Dom0 during those crashes is stable and
never crashed or been restarted…
Any suggestion how to solve this situation?
Downgrade / upgrade to some ”unstable” packages that are known to
be much more stable than those from default Lenny installation?
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
Hi,
first I would recommend you to switch to amd64 kernel at least for dom0.
There were some issues with this 'xenified' Debian kernel version 2.6.26 some
of us were fighting against them [1]. But these seems to be solved in current
kernel.
The configuration option "(dom0-cpus 1)" in /etc/xen/xend-config.sxp
on dom0 could probably solve the issue you are experiencing.
Other things you can do:
- enable DEBUG loglevel in xend-config.sxp with "(loglevel
DEBUG)"
- configure your domU's with option "" -
dom0 will not try to restart it in case domU will crash again
Could you describe your dom0/domU's configuration in more details?
[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=571603
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=542250
--
Peter
