Re: Read Cache, NUMCPUVPS

jeff, : Can someone explain what happened here? Are the numbers : "rolling over"? yup. it rolled over. no problems. do an onstat -z to zero out the profiles if you don't want to look at those negative numbers. it would be nice if this was fixed, as there is no reason that this should be treated as a signed int when we all know that these days doing billions of reads is normal, and in some cases, something that happens every few days or sooner. : Second: : "listener thread: err = -25572: Network driver cannot : bind a name to the port." It then PANICs and the log shows : "cannot wake idle processor." this was a bug that i saw on solaris machines. i either had to wait anywhere from a few to 10 or 15 minutes before i could bring the enigine back up, or bounce the server if i was in a hurry. i believe it was on a solaris 2.5.1 box, with 7.11 or so. i don't remember your platform, but that sounds like the same thing. try bringing the engine up with the 2 cpuvp's fresh after the machine boots, or try waiting around a few minutes after you have shut the engine down before you bring it back up. the problem is simply that the kernel doesn't release the port as quickly and effieciently as it should, and so it is in use when you try to bring the engine up. i don't think that it has anything to do with the second cpuvp as i never saw anything like that, but i could be wrong here. i also ran my machine with 2 cpuvps, and never saw any relation between the two. hope that helps. mickm -- _____________________________________________________________________________ Mickey Mestel mickm@netcom.com -on a beach in thailand to a beautiful, stoned, norwegian woman: "..yeah, it's just another foreign country without ice." -----------------------------------------------------------------------------