Yeah so, I still haven’t gotten the hold on this whole blogging thing. I read some web article a few months back that explained how the whole blogging and public diary thing was on its way out. Summarized, Most people think it’s a cool idea to get up and start speaking to a crowed . ..but they slowly realize that either no one is listening, or that they really don’t have anything to say. I think I fit snuggly into both groups. Ahh what the hell, if anything It’ll be another post for the damn comment spammers to scrape and pillage.

On the home front, I’m the happy father of 2 puppies. One a purebred German Shepherd, the other a Miniature Rat Terrier, respectively named Berlin and Oscar. That by far has consumed much of my time. I’ve been getting involved in Schutzhund, which has been pretty fun so far. I’m getting a lot of pressure to Join the club, but I think the real reason is because I’m one of the only young guys fit enough to be mauled by a dog and to get up and do it again 10 minutes later. It’s a rush when you have a barking, drooling, foaming at the mouth cujo coming toward you. But the thing is . ..it’s all a game to the dog. They are waking their tail the whole time. For me on the other hand, it’s a freakin’ workout!
On the Geek Front, it’s been busy at work.

• Provisioning
• Monitoring
• Virtualization
• Troubleshooting Redhat
• Configuration Management

Provisioning and deploying new or refreshed hardware should be something we already have a handle on at my company, but for reasons I won’t go into now ..we need tighter reigns and a more intuative interface. I tried several solutions such as N1 system manager. However each has their limitations or compromises. In the end I decided to create a simple Web GUI and some backing scripts to handle and validate user input.

Although we are investing in “Enterprise Monitoring” by some big vendor, we in the Unix department still have certain requirement for monitoring and system management. As of lately I’ve been playing around heavily with Zenoss. I’ve been pretty pleased with the Discovery options and simple event monitoring. Again referring to the same frustrations as in the last point, with minimally to no effort, systems will be automagically added to the application and monitored.

We are currently weighing vmware and solaris zones options for virtualization. We are already using zones live in production, but are still trying to feel out the magic configuration for managing this environment. I’m excited about Solaris 10 Update 4, this should allow us to do much more with zones. Certain things in a Shared IP stack have introduced limitations that have caused us to not jump in quite as deep as I would like to have.

Wow, … I have not been very impressed by the error logging or system troubleshooting tools in Rhel4. We recently migrated off of our Sun Sparc systems to Red Hat on x86. Mostly Sun x4100s, but a few X4600s and more on the way. For the most part the transition has been fairly seamless, however after migrating our PRIMARY database over we quickly found some nasty kernel bugs. We’ll just say that was a long week on call. I also had a very difficult time explaining to RedHat support that I can just upload a 64Gb Netdump into the web gui. I’ll give them the benefit of the doubt, maybe they didn’t hear me say GIGABYTE twelves times. Seriously though, their support staff was quick and helpful, I’m just not so crazy about running beta kernels on the system that my Company and my paycheck depend on.

I’ve always been interested in Configuration managment and how other people are doing things in their shops. I’ve really taken a liking to the information available at Infastructures.org. Some I see we can apply to our current setup (based heavily on rdist and a home grown C template processor). I’m trying to find or create a happy marriage of a push/pull system. With the current setup it’s 100% push. Now this push system has worked very well for us, but our infrastructure is about to become very dynamic. Desktops, Virtual Systems, Thin Clients all being up and down, created and destroyed. The only way to manage and support these kind of systems is with a pull component. This is still a work in progress and very much an issue.