I installed the new vCenter Operations Manager 5 about an hour ago. Whoa. I had the pleasure of meeting Kit Colbert, Praveen Kannan, and Martin Klaus last April when they were in town, and I know they, and the rest of the team, have been doing a lot of work on the product. It really shows. Here are my first impressions, literally.
1. The install procedure is very straightforward, in that I was able to get it fully installed without reading any documentation. The product is a vApp now, and comes with two components, the Analytics VM and the UI VM. You need two IPs, and despite the OVF import process prompting for the IPs those IPs need to be in an IP Pool configured for the network. No biggie. You do need some decent CPU, RAM, and disk for this thing to run, though, 4 vCPUs between the two VMs, 16 GB of RAM, 352 GB of disk for the “small” install.
2. It’ll only attach itself to one license, so if you have split your licenses into parts you need to go back into the license portal and combine them. No big deal. I was not able to proactively add the licenses to vCenter, it didn’t like them, but after I registered my vCenter server with vC Ops Manager I was able to. Perhaps the license component of vCenter is extensible, and it does that as part of the registration process.
3. Holy crap, it can send email alerts now. That was my #1 complaint about the older versions. You can also configure the “time remaining” and alert badge thresholds, which is nice. For example, I know it takes my organization at least 90 days to get a new server installed: one month for people to stop telling me that I’m required to use VMs (uhh, these are what VMs run on) and approve the purchase, one month for the vendor to ship, one month for it to be racked & configured & burned in, etc. So I set the time remaining to 90 days (from the default 30).
4. The new notification setup is nice. I can actually get an alarm from something in the standard edition now! It looks pretty simple but I’m okay with that, it looks like it’ll work fine. I’ll mess with this more in the next month. My “upgrade” was basically a fresh install, so it’ll need to learn what normal looks like again. I don’t want it alarming until it does that. Right now, it thinks one of my hosts is in trouble because it’s doing 7.7 Mbps of disk I/O… we’ll let that settle down. That’s actually the power of this type of product, though — the analytics engine that can tell normal from abnormal once it learns what the usual situation looks like.
5. I’m not sure that I like the VMs being children of the physical hosts, but as I think about it that’s exactly the relationship, right? So perhaps it makes sense. The down side will be that I’ll just have to direct my coworkers to use the search functionality when looking for a VM. The whole interface is a lot cleaner, though, and it looks like it’ll be much easier to use than the old versions.
Good job, guys!