As many of you hopefully know there’s a weekly VMTN Community Podcast. VMware’s John Troyer runs the show and usually most of the active VMTN Members join in on the fun. The best shows so far, in my opinion, are the deep technical ones and especially around storage. This evening(for us Europeans at least, wednesday 21:00 GMT +1 !!) Chad Sakac will join to discuss vSphere and iSCSI. For those who don’t know Chad, read this post on VI3.5 and iSCSI and you probably know why you should join in on this podcast. You can dial in, or just listen and have fun in the chatroom. You can ask your questions in the chatroom and one of the panel members will ask them during the call.
Nehalem CPU and TPS on vSphere
As I wrote a while ago when you enable virtualized MMU for your virtual machine it enables Large Pages and Large Pages don’t get “TPS’ed”. The article I wrote was specifically related to AMD cause it was the only platform at the moment for which enhanced memory techniques where used. (AMD RVI!) As of vSphere 4.0 Intel EPT is also fully utilized. As expected this leads to the same “issue” as with AMD, no TPS when you enable vMMU. VMTN Community User MCWill reported this here. I wanted to specifically point this topic out to you because of the excellent replies from Kichaonline and Rajesh Venkatasubramanian. It’s worth reading the full topic if you want to get a good understanding of TPS/Virtualized MMU.
A small correction — we are currently investigating ways to fix the high memory usage issue also. Regarding TPS, as noted earlier this shoud not lead to any performance degradation. When a 2M guest memory region is backed with a machine large page, VMkernel installs page sharing hints for the 512 small (4K) pages in the region. If the system gets overcommitted at a later point, the machine large page will be broken into small pages and previously installed page sharing hints helps to quickly share the broken down small pages. So low TPS numbers when a system is undercommitted does not mean that we won’t reap benefits out of TPS when machine gets overcommitted.
Disk latency and esxtop
We just had a very good and interesting VMTN Podcast on virtualized MS SQL performance and best practices. One of the questions was about disk performance. Hemant Gaidhan talked about esxtop and how to discover possible performance issues, and specifically mentioned latency. I’ve never really looked into this section of esxtop and did a quick search and of course the “Interpreting esxtop Statistics” answered which counters to watch and what each counter represents:
Section 4.2.2 Latency Statistics
This group of counters report latency values measured at three different points in the ESX storage stack. In the context of the figure below, the latency counters in esxtop report the Guest, ESX Kernel and Device latencies. These are under the labels GAVG, KAVG and DAVG, respectively. Note that GAVG is the sum of DAVG and KAVG counters.
I recommend reading the rest of the 4.2.2 section to anyone looking for more indepth info on esxtop and storage performance. Also read page 14/15 of Hemant’s document on SQL Server performance/best practices. Another great read and tip from Hemant was the “Scalable Storage Performance” whitepaper.
VMTN Podcast: Join in on the fun!
This evening(well at least for us Europeans) there will be another edition of the VMTN Podcast. More and more people are joining the weekly podcast live(Live audio stream and chat) and more people are downloading it every single week, jump on board while you still can! This week John Troyer asked VMware’s Hemant Gaidhan to join us and talk about SQL Server workloads in a virtualized environment and of course best practices.
Hemant will be our guest on the VMware Communities Roundtable podcast tomorrow, Wednesday, April 1, at noon PDT / 3pm EDT. I believe Europe and the US are now back in sync with respect to summer time, but you can always check a time zone calculator.
You can dial in, join the chat, and get streaming audio here. Your homework assignment is read Hemant’s paper and come with questions.
VMworld, the wrap-up
VMworld Europe 2009 came to an end on Thursday. I finally traveled back on Friday, after a 5 hour delay.
I’m sorry I haven’t been able to write any technical content but you can’t imagine how busy I’ve been. I’ve gathered all useful links and added them to this post. But just to give you an idea what I was up to during VMworld:
Arrived on Sunday at 15:00, dropped luggage at hotel and went to the Hands-on Labs. Started unpacking the thin-clients and helping with the lab environment till 20:00. Changed clothes and went to the Veeam party.
Monday, 08:30. Preparing the labs again. Testing lab manuals, setting up thin clients and solving SAN issues. Finished at 21:00 and went directly to the Dutch VMUG Meet and Greet.
Tuesday, 08:30. Back to the labs again and do some final testing. At 10:30 first people entered the labs and I had Community Booth duty the rest of the day till 19:30. Quickly changed clothes and back to the VMware Benelux party.
Wednesday, 08:30. Watched the keynotes and hurried off to the community booth again. Helped a couple of minutes at the labs but mainly booth duty. Went to the hotel at 18:30 to change clothes for the VMworld Party Cloud9.
Thursday, 08:30. Lab duty for the first couple op hours. Booth duty the rest of the day. Couldn’t feel my feet anymore after a couple of hours. When the solution exchange closed at 16:00 we needed to put everything in the original boxes again. All thin-clients / monitors / keyboards etc. 19:30, done, change clothes and went out for a decent meal… actually the first decent meal this week.
This, was just a quick write-up. But you can imagine that most VMware employees have been running around all day long, no time to eat / drink / sit down. Talking to people all day long, answering questions or just random chit-chat, there’s even no time to reflect and digest all the announcements. Keep in mind that even weeks before VMworld people are doing preparations. For instance the VI-Toolkit lab had been tested by at least 6 people two week before VMworld. When I visited VMworld as an “attendee” last year I honestly didn’t have a clue how much work an event like this really is.
For the Hands-on Labs: At least 1000 attendees visited each day. Yes a THOUSAND at least. I really loved the format of the labs, the open floor instead of the closed lab sections was also a great idea! I would have never expected that there would have been queues, but there were… Sometimes even over 20 people were waiting just to get in. Definitely a big compliment to all Hands-on Labs staff, they/we did a great job.
And let’s not forget the team that organized this event, great job! Thanks!!!
I hope I will be able to attend VMworld 2009 in San Francisco. Over 15.000 attendees are expected at the Moscone Center: