I think that some info needs to be cleared up. sysctl -w kernel.hz_timer=1 will DISABLE the timer patch. sysctl -w kernel.hz_timer=0 will ENABLE the timer patch. The best way I have found to check this(only when running under VM) is to use the CP trace facility (tip from the Boeblingen lab). Start a CP trace on an idle linux guest with this command: #cp cpu all trace ext 1004 run If the timer patch is not enabled, your console will be filled with 100 messages every second(multiplied by the number of processors defined to your guest). If the patch is enabled, it will be much less than that, perhaps only a few per second. Stop the trace with this command: #cp cpu all trace end We run one IFL and with the hz_timer=1 setting we get 100 hundred messages, with hz_timer=0 we get 3-10 messages approx. each second.
Kevin Ellsperman Payless ShoeSource, Inc. 785-559-6326
