Re: 3390-9 performance considerations

[Tom Rae (WFF)]

The only “gotcha” I have found in researching this is the possibility of increased device queuing on the operating system side. Since VM does not support PAV yet you end up being able to do less concurrent I/O due to device control block queuing. If your workloads do not require a lot of concurrent I/O then using larger volumes should not be a problem.   Tom Rae Senior Director, Technical Services Western Canada Loblaw Companies Limited Information Systems Division From: VM/ESA and z/VM Discussions [mailto:[EMAIL PROTECTED] On Behalf Of Carlos A Bodra Sent: October 18, 2005 10:27 To: VMESA-L Discusion List Subject: Re: 3390-9 performance considerations   3390-9 or 3390-27. Doesn't matter. All I/O operations ends when data arrives in large cache memory. I have some 3390-0 volumes and no performance problems. Carlos Alberto Bodra S/390 System Programmer Sao Paulo - Brazil     From: VM/ESA and z/VM Discussions [mailto:[EMAIL PROTECTED] On Behalf Of Kurt Nelson Sent: Tuesday, October 18, 2005 12:31 PM To: VMESA-L@LISTSERV.UARK.EDU Subject: 3390-9 performance considerations We're running VM 5.1 with SLES9 and Oracle, currently using SHARK dasd emulating 3390-3. Our dasd folks want to know if we have any problems with the SHARK dasd emulating 3390-9. I believe everything supports that ok, but I'm wondering what perfomance considerations there might be both pro and con to using 3390-9 emulation? I would appreciate any information/experiences that anybody might have.   Thank You. Yahoo! Music Unlimited - Access over 1 million songs. Try it free.