And when calls come into the service stream, our first response is something along the lines of: Greetings The most noted cause of the condition you are referring to here is related to the resources your system has, compared to the resources given to your Linux guest(s). Next time this occurs a: CP INDICATE QUEUES EXPANDED (ind q exp) command from a class E user could prove this. If the user/guest is in the eligible (E) list, that would indicate that the guest is waiting on resources. This would also explain why the guest can free itself. Enough resources become available for the guest to run and they are moved out of the eligible list. Other (not so common problems) can reside in: 1) MTU size mismatches between network, guest(s) and system. 2) Quick Disp settings.
Please see http://www.vm.ibm.com/perf/ and more notable within this page: http://www.vm.ibm.com/perf/tips/linuxper.html for more information. Best Regards, Kurt Acker Rob van der Heij <[EMAIL PROTECTED]> Sent by: Linux on 390 Port <[EMAIL PROTECTED]> 08/12/2003 04:01 AM Please respond to Linux on 390 Port To: [EMAIL PROTECTED] cc: Subject: Re: Virtual machinnes freeze > Hi all, today we experienced some strange behavior with our vm, we have 12 > linux virtual machinnes running, tcpip, and monwrite, the problem is that > suddendly one linux vm freeze without reasson, with low cpu utilization > (0-10%), no pagging, no load, it just freeze, when we acces trough vm This is the eligble list. Most likely your total storage requirements of all in-queue virtual machines is beyond what CP considers healthy, so one or more of them are kept aside until there is room. The defaults in z/VM and the way Linux virtual machines interact with z/VM make this process erratic. Look in the Performance Redbook for the full explanation and how to get out of that situation (and avoid any suggestions that let you fix this in 1 minute). http://publib-b.boulder.ibm.com/Redbooks.nsf/RedbookAbstracts/sg246926.html? Open Rob
