After upgrading some of our clusters, I noticed that some metrics were not being sent out or updated. It took about 10 minutes for the following metrics to be included in the XML output and displayed in the web-frontend:
load_fifteen, proc_run, cpu_aidle, cpu_speed, mtu, swap_total, load_five, disk_total, cpu_num, mem_total, sys_clock, cpu_speed, machine_type, disk_total, os_release, boottime & os_name. Even restarting a gmond in the cluster did not seem to force a re-broadcast, and I thought that was how it was designed. This didn't happen on every cluster or even on every node in the clusters that it did happen on. I also noticed in the webfrontend, the list of metrics that can be viewed and sorted on, in the cluster view, no longer include gmond_started & [last] reported. Why were these removed? I briefly looked in header.php but didn't see anything obvious. ~Jason -- /------------------------------------------------------------------\ | Jason A. Smith Email: [EMAIL PROTECTED] | | Atlas Computing Facility, Bldg. 510M Phone: (631)344-4226 | | Brookhaven National Lab, P.O. Box 5000 Fax: (631)344-7616 | | Upton, NY 11973-5000 | \------------------------------------------------------------------/
