Bug#464324: cman: Cman stops working when starting/rebooting a xen dom0
Package: cman
Version: 1.03.00-2
Severity: important
The following problem only exists with ccs+cman running on a xen server
(dom0).
Each time I reboot a xen server of the cluster, CMAN stops working
during the boot process as soon as xend starts. It seems that it is
caused by xend which does something with the ethernet settings. I have
found a report which is quite close to the situation I'm facing to:
http://sourceware.org/cluster/faq.html#xenclustererrors
As a workaround, I have added a new init script which is launched before
ccs and cman. This init script just builds the required xen bridges:
/etc/xen/scripts/network-bridge start netdev=$IF_DEV
I have disabled the bridge builds in xend like suggested on the above url:
"network-script /bin/true" in /etc/xen/xend-config.sxp).
That way, I have no more errors from CMAN when starting/rebooting a
cluster node. I wonder if this is the best way to get rid of this
problem, according to the debian use and wont.
I have also tested to build the bridges (still using "network-bridge")
directly from /etc/network/interfaces (using post-up) but for some
reasons it didn't work.
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: amd64 (x86_64)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-5-xen-amd64
Locale: LANG=fr_FR@euro, LC_CTYPE=fr_FR@euro (charmap=ISO-8859-15)
Versions of packages cman depends on:
ii ccs 1.03.00-2 Cluster configuration system
ii libc6 2.3.6.ds1-13etch4 GNU C Library: Shared libraries
cman recommends no packages.
-- no debconf information
Reply to: