Several problems have been encountered in the field after a person runs 'shutil -i' but not 'shutil -s /etc/cluster.xml'. Since, practically speaking, they should always be run together, the proposal is to have 'shutil -i' also store /etc/cluster.xml in the configuration section of shared storage. This will prevent shared storage with a blank cluster configuration, which causes clulockd and clusvcmgrd to complain: Sep 8 18:14:15 foo01 clulockd[2342]: <emerg> Could not determine local member ID! Sep 8 18:14:15 foo01 cluquorumd[2246]: <crit> Lock manager missing; restarting! Sep 8 18:14:15 foo01 clulockd[2344]: <emerg> Could not determine local member ID! Sep 8 18:14:15 foo01 cluquorumd[2246]: <crit> Lock manager missing; restarting! This does not preclude the user from overwriting with another cluster.xml; it simply prevents the common case of initialization without storing configuration info.
Created attachment 103845 [details] Patch implementing requested behavior.
Verified the above test against: clumanager-1.2.22-2 Ready for RHEL3-U4.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2004-491.html