Description of problem: It would really be nice for ccsd users if there was an actual "ccsd update the cluster.conf file" command. This way it could actually tell you wheither or not your changes made it out to all the nodes or not. Currently one has to find ccsd's pid, send it a signal, check the SYSLOGs on all nodes, and then check the cluster.conf files on all nodes to see if it really worked or not.
I will make said script for you - it involves a couple other small things (like automatically notifying cman or gulm of the changes)... For now, just do #!/bin/bash killall -HUP ccsd if[ -e /etc/cluster/cluser.conf-rej]; then echo "Update failed. Consult /var/log/messages for the reason."; echo " Your changes have been saved in /etc/cluster/cluster.conf-rej"; else echo "Update succeeded." fi
Updates with the proper version and component name.
Updates with the proper version and component name. Again, just love out tools.
ccs_tool update [new_cluster.conf] will update and report results It does not yet tell cman to update.
when the update fails (for instance if ccsd dies on another node) this script tells you to search the logs for your answer why, it'd be nice if it reported that error so one doens't have to search n number of logs. [root@morph-04 root]# ccs_tool update Cluster update failed. Update failed on remote nodes. Check their logs for clues.
update tool is available. Openning new bug to reflect remaining gripe.
verified availability. :)