by
adam baum
Day Three
I only did a half day today since I had to get to checkout of the hotel and get to the airport. I did manage to attend two troubleshooting sessions. I've noticed a pattern with these sessions: they are either too short or too complicated for the time allotted. Both of the sessions quickly reviewed the ESX boot process and diverged from there. The first session focused on where to find info. The presenter threw in various tasty tidbits of info, but he pretty much covered where all the log files are. One thing I didn't know is that on the VC side of things, VM loves to use the various temp directories. If you need info, look for a temp directory (don't forget ALLUSERS profile). Chances are there are log files in it. In 30 words or less, here are the main things to look at:
Boot process
VI Client
DNS
Console
esxtop
vm-support -s
vmotion
HA
All of these either have log files or can be manipulated in some manner to provide useful info.
The second troubleshooting session was called "Tips for Troubleshooting ESX Server 3.x Faults". After detailing the different classifications of problems, the presenter had us debugging PSOD screens. yeah right...I don't think I'm going to learn how to do this in one hour. The funny part of this is that the error used was a Machine Check Exception. Once you get down to the final codes for MCE, you'll need to reference the CPU architecture guide provided by the CPU manfacturer. I'm sorry, but I don't have the time, nor the skill to sift through an engineering document of this type. This reminds me of yesterday's sessions where we told to download a source code header file to research error codes. I'm not lazy and I do like a good puzzle, but I pay good money for support and maintenance. If I get that bogged down, I'm making a phone call.
My last session was based on VMware's vision and strategy for datacenter mgmt. It was actually decent. The only thing I didn't like was that they tried to show how their products mapped into ITIL functions. I can't stand how every vendor now thinks they offer an ITIL solution. It reminded of how every storage vendor had an ILM solution when Sarbanes-Oxley first came out. Note to vendors: ITIL is all about processes, not products. Your product is not an ITIL solution.
Barring the ITIL references, the presenter did a good job of explaining how they see the Dunes acquisition panning out. If you take the standard VM architecture slides, add in process automation/control and delegated administration at the top.
The best way to look at process automation and control is to compare them to document mgmt workflows. There is check out/check in, there are approvals, there are modifications, there are archives, deletions, etc. With the Dunes acquisition, VM will be able to provide a workflow framework. Think of self-service provisioning based on rules you setup. Think of Lab Manager capabilities delegated to specific groups. These capabilities should free up the VM admin so that he (or she) can spend more time proactively maintaining the system. If I'm not making sense, please review the session materials that should be available on VM's website next week.
Final thoughts....overall, I think it was a good show. I learned a lot, met some interesting people, and made some good contacts. I think I met my goals that I laid out in first blog entry. Was it worth it financially to my organization? I don't know. The majority of the info I took away will soon be available from the posted sessions. I think the networking may endup what pays for this trip. In particular, I will be joining a user group focused on government and educational entities. Since we have no competitors in the traditional sense, we tend to be more open with the info we share.
Going off track a bit...did anyone else notice that the show was using eval versions of Camtasia to record the sessions? Nice.
I am going to revisit a few of these topics over the next few days once I have time to digest all that I've seen and heard. In particular, I want to take another look at Akorri's BalancePoint and VM's Site Recovery Manager. I'll also comment on VMWorld's use of customer provided content and how well it worked (or didn't work).
(Note: You must be logged in to post a comment.)