Forgot your password?
New user?
Document Actions
Tracedump and memorydump of a clusternode
I all cases when a cluster makes problems and should be analyzed further it is very important to get a kernelstacktrace and memorydump as described below
Prerequesits
If the whole cluster is freezed and you cannot get a shell issue the commands via the fenceackserver. Therefore connect via telnet to port 12242 and login as configured in your /etc/cluster/cluster.conf. Then got to shell mode and execute this commands as prerequesit
mkdir /proc mount -t proc proc /proc
NOTICE
This will be changed in future versions and replaced by just issueing the sysreport command on the fenceacksv command shell.
Analysis
Login to any clusternode either via fenceacksv or via shell and issue the following commands.
echo t > /proc/sysrq-trigger echo m > /proc/sysrq-trigger
That's it now collect the syslog output and there you go.
Follow @OpenSharedroot on twitter