Question RE... onstat -g spi - tcb_queue_lock

Good Morning, IDS 11.50.FC5 O/S AIX 5.3 I have a system that sometimes runs into perf problems but they are sporadic. While looking into this I found that the number of average loop / wait is more than twice as high as in any of my other systems for tcb_queue_lock. The first few lines of "onstat -g spi" look like this: Num Waits Num Loops Avg Loop/Wait Name 3 8 2.67 mtcb lock 41403011 2845842149 68.74 mtcb sleeping_lock 307 661934 2156.14 mtcb tcb_queue_lock Could this possibly be at least a partial cause of my performance problem? If so how can I bring down that number? TIA, Dan

Hey Dan, 41403011 2845842149 68.74 mtcb sleeping_lock 307 661934 2156.14 mtcb tcb_queue_lock I would say the sleep_lock has a very good metric for such a hot resource, its avg wait time is only 68.74. the tcb_queue_lock probably can be discounted (since its not a hot resource) even though the wait_time is what I would consider very high. Mark