>It depends on what you mean with "performance testing" Some people >try to measure maximum throughput of a single server. Though this may >seem easy to do, it has little business value for Linux on z/VM. In an >environment where resources are shared and servers are utilized only >part of the time, measuring maximum throughput does not tell you a >lot. 10 virtual machines at 10% utilization behave completely >different from one server at 100% utilization.
I agree with you on this measuring maximum throughput. But I think the dedicated CPU makes sense if we want to perform the benchmark compared to distributed platform. And it's also meaningful if we try to perform the stress verification test. Generally, we won't want each test result would be different due to other guest machines' influence. I think dedicate CPU to that Linux can help to eliminate the impact from other guest machines. >If your ultimate goal is to run Linux in an LPAR with dedicated IFL, >then your measurements on z/VM with a dedicated IFL (or shared only >with yourself) will be pretty close. But very few run it like that. Yes. But I prefer to running the zLinux under zVM. >We measure resource usage of the virtual machine and divide that by >the number of transactions. You do that at different levels of >utilization to understand how scalable the application is. You will >often find that efficiency gets worse at very high utilization (lock >contention, for example) and also often at low utilization (idle load, >polling, etc). I don't understand how you can know the scalable of the application via this approach. Can you get the accurate application scalability? It's estimated pro rata? Maybe my questions are not so 'professional' for a tester... Best Regards Anson Y ________________________________ 发件人: Rob van der Heij <rvdh...@velocitysoftware.com> 收件人: IBMVM@LISTSERV.UARK.EDU 发送日期: 2010/4/13 (周二) 4:57:16 下午 主 题: Re: RE: LINUX on IFL If your ultimate goal is to run Linux in an LPAR with dedicated IFL, then your measurements on z/VM with a dedicated IFL (or shared only with yourself) will be pretty close. But very few run it like that. > It's an interesting topic! If you were going to test the application > performance on zLinux, what will you do? Install Linux on LPAR directly? We measure resource usage of the virtual machine and divide that by the number of transactions. You do that at different levels of utilization to understand how scalable the application is. You will often find that efficiency gets worse at very high utilization (lock contention, for example) and also often at low utilization (idle load, polling, etc). Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/ On Tue, Apr 13, 2010 at 10:29 AM, Anson <yeal_c...@yahoo.com.cn> wrote: > Do you mean the zLinux running under zVM couldn't be used for performance > testing purpose, even we dedicate one CPU to this guest machine? It depends on what you mean with "performance testing" Some people try to measure maximum throughput of a single server. Though this may seem easy to do, it has little business value for Linux on z/VM. In an environment where resources are shared and servers are utilized only part of the time, measuring maximum throughput does not tell you a lot. 10 virtual machines at 10% utilization behave completely different from one server at 100% utilization. I believe with Linux on z/VM your objective should be the most efficient (cheapest) way to deliver the service within the SLA. Running a single virtual machine at 100% does not provide much insight in this area. > From my point of view, I think it's reasonable... We just don't want to > create a new LPAR with dedicate CPU.