ICMP processing statistics appear in the output of the sm_tpmgr utility and will be similar to the following:
ICMP Accessor Interface: avg_late_polling = 0.250944942235947 bytesPerPing = 64 gets_causing_request_percentage = 0 gets_from_cache_percentage = 100 icmpNumberOfPolls = 27109905 icmpNumberOfResponse = 1252990 icmpPollerTimeSkew = 0.250945 icmpStartTime = March 14, 2009 5:52:58 PM EST max_active_processing_time = 0 max_get_time = 0 max_idle_processing_time = 0 max_late_polling = 0 max_lock_wait = 0 min_get_time = 1000 num_other_failures = 0 num_threads = 10 num_timeouts = 0 operation_size = 1 periodic_gets_per_second = 0 total_active_poll_actions = 0 total_get_nexts = 0 total_gets_from_cache = 30 total_instrumentation_get_requests = 0 total_on_demand_gets = 0 total_periodic_gets = 0 total_poll_actions = 0 total_repos_gets = 30
Look for these statistics to evaluate performance:
-
max_get_time, min_get_time: The maximum, and minimum duration (seconds) of a ping cycle.
-
avg_late_polling: The amount of time (seconds) that the ICMP pinger is falling behind. If the value is negative, the pinger is ahead of schedule.