Yes, This was brought up in a recent lab/class in Poughkeepsie. Currently under VM including VM 5.2 being released this constraint still exists. the SIE instruction as stated is being emulated in software in a second level VM system. I think Denise said they where working on improving this but currently and in 5.2 you will see degration in performance and we did witness it in the class.
William 'Doug' Carroll Mainframe Systems Engineer II Global Technology Infrastructure ECS Mainframe Operating System Services Explore IT, build IT, exploit IT - Creating excellence through teamwork. (614) 213-4954 Office (877) 899-1697 Pager (614) 244-9897 Fax http://www.bankone.com http://www.jpmchase.com James Melin <[EMAIL PROTECTED] To: [email protected] epin.mn.us> cc: Sent by: Linux on Subject: Re: Poor performance running VM under VM in a DR exercise 390 Port <[EMAIL PROTECTED] st.edu> 11/09/2005 02:06 PM Please respond to Linux on 390 Port I spoke to an IBM expert on things WebSphere on z/Series and he had this to say: _________________________________ You will experience much degraded performance when you are running Linux under VM under VM. My understanding of this is that it is caused by running the SIE (start interpretive execution) hardware instruction in software. VM uses the SIE instruction to run guest operating systems like Linux. When VM is itself running interpreted (2nd-level) then when it uses the SIE to run Linux that SIE is running in software emulation. So it's not just your WebSphere that is slow, but all of Linux, and any other guests running under the 2nd-level VM. You're probably not noticing the slowdown in Linux since Linux is quite "light" and doesn't really put much of a load on the system. But WebSphere will, and you'll notice the performance degradation. __________________________________ We are running VM under VM for the convenience of the DR Vendor. Rather than have to re-create our VM environment for us with different device addressing.... A couple questions : Any way to prevent third level guests from doing SIE in emulation? When we do our z/VM & z/Linux restores, the process we undertake is to do it from the 'starter system'. We restore to the real disk devices on the recovery system. The process of volume restore for Linux changes the Volser, so when our VM runs, the devices are at the address they've been told to be at in the VM directory and the labels have been changed to be what labels VM expects. The only real thing that needs to be resolved is recovery system device address definitions vs our 'native' device address definitions. Is it better to re-map our VM configuration to match the device addresses of the recovery hardware, or have the DR vendor create guests in the primary VM that match my Linux definitions? ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
