VMware Update Manager for Windows VMs

[ Note: yet another technical post ! I just can't resist it... ]

I had never taken the time to look into Update Manager. Today, I decided to dive in and try it out with a Windows Template I'm building.

It struck me that the update process (actually it's called 'remediate') took so long. Nothing could be seen on the server console and no CPU was utilized. Strange.

After a couple of minutes waiting, I went back to the VM console and saw a CD drive mapped. I should have known! Windows updates are deployed the same way the VMware tools are: via a virtual CD (an ISO file mounted from the virtual center server). Indeed on the VC server, I found an ISO with the exact patches I selected to be installed on the VM.

That's why the first phase of the remediate process took so long: it was preparing the ISO.

On a sidenote: This ISO can as well be used for physical machines that need to be patched! Just copy the ISO file, burn it on a CD. The only thing missing still is the so-called Update Manager guest agent which is installed on the VM to be patched. It seems that 'vum-launcher.exe' is the thing that does most of the work. Did anyone test this out already?

[Update] VMware Bug: Waiting for a patch

I was right in expecting that a host reboot would not be required to intall the patch.
I was wrong to think this means I did not have to move my VMs away or shut them down. Maintenance mode is required to install it. Lame!

Fortunately, I still have some 3.0 servers around.

VMware Bug: Waiting for a patch...

Ok, I couldn't resist... I had to get my opinion out on the 'by-now-famous' bug in the ESX 3.5u2 hypervisor.

This morning, I was worried. I imagined having to shut down several updated ESX servers hosting more than 100 VMs, patching the servers and bringing the VMs up again.

Looking around and discussing with Tim Jacobs, I looked back at one the posts he refers to. It struck me that the error is logged in '/var/log/vmware/hostd.log'. This means that it is the host management process that is logging the error. To me, it would make sense if the VMkernel doesn't care about licensing and just does its job.

As a consequence, it must be possible for VMware to create a patch that does not require a host reboot.

This afternoon, I look forward to such a patch.

Up to something different ...

According to Google Analytics, this blog has between 25 and 50 visitors per day, with over 70% from search engines. Feedburner tells me that there are little over 50 feed subscriptions, which I find nice. The small Google banner on the right has generated around $20 in five months or so. All these things make me happy. This might soon be over.

Virtualization and VMware products in particular are nice to work with and blog about. Especially when it comes to capacity planning and analysis, a lot is still to be discussed. But not by me.

I made a jump (content-wise) about 5 years ago, going from an academic research group to the IT industry. The jump I'm about to make in a couple of weeks is probably even bigger. Probably starting in a project management function, with the intention of getting into BPM, Service Management or who knows what...


Why Management Consulting?

Because of the money? No
Because of the challenge? Probably
Because it interests me? Sure enough.
Because it's related to what really keeps me busy. Sure!

But also because I do not see myself implementing VMware infrastructure products (or any other product for that matter) for 5 more years. It is not the thing I have in mind for my older days. Is management consulting the answer? Perhaps, we'll see.


Why not IT?

There are some possible answers:

1) What really interests me with products like VMware and such are the foundation layers, the theory behind, the statistics, the stories, etc. This probably stems from my theoritical background. Suffice it to say I was most interested in the talk of Irfan at VMworld Europe 2008.
This is one of the main reasons I was attracted to VMware in the first place, just take a look at some of the earlier posts to see that people at VMware are publishing papers in scientific journals.

2) What I liked a lot in past projects is what we call the 'design phase'. Discussing with a client about requirements, expectations, boundary conditions etc. and coming up with a good compromise that is cost-effective. This requires thinking and communication, two things I miss a lot in the 'implementation phase' of a project.

But then again, looking at the Belgian context, making a design for 3 ESX servers and 2 VLANs isn't really that exciting. We simply don't have that many large corporations in Belgium and the cake has to be shared with other consulting firms.

3) When designing or implementing a software product, one is bound to what is deliverd by the vendor. It is very frustrating to deploy a solution, only to find out that the software contains a bug that only pops up under the specific circumstances at the customer's site. At best, you can get a hotfix or patch from the vendor to cover this up, but bottom line, we depend on the quality of work of others.

By the way, coming from a Linux background, I'm convinced that this last argument is an important one in the discussion about open source versus closed source software. Agreed, most of us are not capable of coding our own software, but if required we could hire an independant developer to fix our bug when the vendor does not support us!

Did I have bad experiences with VMware? With Citrix? Or with any other product? Yes I had, with most of them. Think about some vendors not giving support to applications that run in a virtualized environment, think about the famous VMotion bug that has caused a client of us to have a lot of issues with an SQL server, think about so many other things that cause us to spend 80% of the project time getting the 20% of the product under control.

Update: I typed most of this letter last week, could I possibly know that today (August 12, 2008) would be called "D-Day for VMware" by Tim Jacobs and even worse by others? Ironically, I upgraded most of the servers at a client site yesterday (11th), only to find out that a bug causes major havoc as of today!

4) I like to talk to people, think, brainstorm, discuss on the blackboard, etc. Feeling synergie when people get together. Giving training (as I used to) comes close, but is often too much one-direction. Workshops (as I do often now) is good, but doing the same thing 10 times in a row is not my intention.

5) Technology changes really fast. At first, I found this exciting. Nowadays, I sometimes think it has become a burden. Don't understand me wrong: I have nothing against new features and clever products and I'm the first to check out what's new. The expectations of some clients, however, are such that sometimes you're expected to know all about these things before they are released and that you know these things by heart.

The above deals with the content of the IT work. I have a lot to say about the form as well (how project consulting is misunderstood and how people are resources, rather than assets). That is an intirely different discussion.


Back to the Future

What about the future? Well, as my current interests and future work will be about different things, I'm still unsure as to what I'll do with this blog.

Somehow I like blogging about what keeps me busy. I am co-author of a (dutch) blog (http://www.slimmerwerken.be) and what does not fit there usually ends up on my personal (dutch) blog (http://choose2live.blogspot.com).

What holds me back to continue verbeiren.blogspot.com is the fact that in IT, nobody really cares if your English text is not perfectly written without spelling or grammar mistakes. When writing about business processes, communication itself or personal productivity this may not hold true.

So the first question is whether I should at all blog about my future activities (in English). The second question naturally follows: should I keep this blog (and change the subject/content), or stop it here and create a different one.

If you have any advice, please let me know in the comments.

VMware, backup and VSS

People looking for information concerning backup and the latest features included in ESX 3.5u2 should take a look at this article by Tim Jacobs.Not only does it introduce VSS, but it also details why VSS support for VMware backups is a good thing.

Keep up the good work, Tim!