> Hello all,
> Can anyone please suggest a legal way out to figure out if the > link between the node of a cluster and a master host (where gmetad > resides) is lost? It's relatively easy to find out this fact being > on the master host - if we don't see responses from a particular node > for some period of time, then we can consider it to be dead. (I know > that this idea may not work in some cases, but it's ok for me) > But how one can do the same thing being on the node? The reason why > I'm asking this, is because I need to perform some cleanup operations > on the node in case of connectivity failure (for example network > interface failure). > Artem Pervin I think I've found a suitable solution. I can start gmond with "-d 2" option in order to log it's activity and then grep "heartbeat" messages. If gmond sends heartbeat messages without errors, then the node is OK, otherwise if gmond reports something like > sent message 'heartbeat' of length 8 with 1 errors during some period of time, then the node can be consider dead. Artem. ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
