A little late, but here nevertheless.
Breakfast for today was a delicious Mel’s Diner skirt steak with eggs and a fresh squeezed OJ. Damn. So good.
No keynote today so it was straight to a full day of sessions.
First session of the day was INF6394-GD vSphere Lifecycle with Salil Suri and Brian Graf. The focus was really around issues that the group had around the installation/maintenance of vSphere and it’s components. The discussion focused more on issues with host profiles and why people weren’t using them. Toward the end we discussed how packages were deployed and applied (VUM and the like). One of the feedback items so provided was to deprecate out the storage portion of host profiles to a separate section and to make error management easier. Another point I made was that VMware should do more with OEMs to integrate firmware maintenance as part of VUM. I also met a guy who was using kickstart in a very similar manner, I gave him a card and hopefully can sync up with him after the show.
Nect up was INF5729 – vSphere multi-processor FT
I have a very vested interest in FT, we run possibly a dozen FT machines and hit some critical issues last year. It took 9 months working with engineering to get it fully resolved, and unfortunately I got pretty terse with the engineer, who was also the speaker for this presentation.
The presentation recapped the SMP FT technology, limits (64gb ram, 4vCPU), 10gb network, and then performance benchmarks. Apparently SMP FT is optimized for Haswell cpu, something on the order of 10-12% better performance than the prior chip. I did ask if EVC mode would cause any issues and the answer was no. Most of the tests that they ran showed fairy good comparisons to non-ft except for the heavy sql workload. A white paper on the tests and how to recreate will be coming out soon.
Lastly I hung around to chat with Jim (the engineer that I gave so much s*** to). I did apologize for giving him such a hard time and he seemed to not hold a grudge (probably just happy to get rid of me).
Next on the day’s agenda was lunch. The first two days of the week, the Intercontinental would have a lunch spread for those who were having meetings. The options today weren’t as good as the previous but better then Moscone. I believe I had a roast beef sandwich with a small salad.
SDDC4593 was after lunch, it was a panel of top vbloggers including Duncan Epping, Scott Lowe, Chris Wahl, Chad Sakac and Rick Scherer. Questions were often pretty light and responses introspective or funny. Hearing about each panelists home lab was cool but the most important thing that struck home was that Duncan said that the area that he had to work on the most was public speaking. This was definitely the most humorous session. A funny thing was that William Lam (one of the top bloggers now) was sitting behind me.
Last session for the day was STO6555-GD: Meet the Site Recovery Manager Engineers. One other attendee chuckled and told me that it was kinda cool to be in a session with the presenter (myself) of another session. Hopefully he wasn’t one of the low ratings I received.
Super Duper for dinner, some drinks at s hotel and then Off to the party (which was really good)
One note for the day, I was able to approach a problem from three different angles today.When you have SRM and your vCenters are in linked mode, if you do a search for a can that is also protected, you will see the VUM and the placeholder vm. They both have the same icon and if you search again the order may be different.
I brought this up with the Web Client PM yesterday and he said it was the SRM team’s issue. Today I brought it with three SRM engineers who noted it down though he mentioned it was on the UI side. When I was walking to Super Duper I ran into the whole SRM UI team from Bulgaria (whom I met before). He said it may be the client but admitted it may me them and would look into it. Not sure how much of a commitment that is but I at least brought it up to all of the relevant parties. Only at VMworld.