With VMworld Europe in full swing we can expect some product announcements and press releases happing every day.
VMware ESX 3i Embedded will be available from at least 4 major vendors “real soon now”. At least that’s the word on the street. I’ve definitely heard similar things from my vendor contacts.
I’m very excited about this since it will be great to use in some of our regional sites. It’s tough to get disparate hardware in and have to juggle configs around to get on that new hardware remotely. This should really smooth things out.
Here are some announcements in that arena:
Fujitsu Siemens announces immediate availability: http://www.vmware.com/company/news/releases/fsc_vmworld.html
HP announces March 31 availability:
http://www.vmware.com/company/news/releases/hp_vmworld.html. This comes with the announcment that HP Systems Insight Manager 5.2 will support ESX 3i. HP SIM 5.2 is already available here: http://h18004.www1.hp.com/products/servers/management/hpsim/index.html
Dell announces early April availability:
IBM is only included here as part of the “available within 60 days”:
http://www.vmware.com/company/news/releases/3i_rollup_vmworld.html, however a search of their site shows they are well along in preparing to support 3i.
Welcome to a giant paradigm shift in computing. The future is warming up on the runway and should launch in the next couple weeks.
VMware ESX 3i is ESX 3.5 with the Console OS (Service Console, COS, what have you) removed. Gone. Kaput. No more logging in to fix things. No more backup or management agents running on the ESX host. Check out VMware’s product page and datasheet for 3i.
Does that sound like big deal to you? If not, it sounds like you’re ready to move on down the road. Just bump your VirtualCenter load to 2.5 and get after it. However, if this move leaves you confused about how to handle backups, managment agents, or other services let’s cover the pain points and how VMware expects you to deal with the change.
- Custom Scripts: Do you use bash or perl scripts to gather performance data, perform some process related procedures, or other custom sequential operations? Expect to port these to either the new remote CLI (Command Line Interface) or get used to using one of the other APIs on a remote system.
- Management Agents: If you use servers from major manufacturers you may use their management agents to report hardware and software events to a centralized management server. Those agents are now useless for monitoring a 3i system. The good news: An industry standard CIM interface is used to expose all of those things you’d expect to see. The bad news: We’ll be waiting on the manufacturers to update their management platforms and they’re notoriously bad about hitting launch dates with supporting agents. In a big shop this will either force the use of the full version of 3.5 as a transitional stop-gap or slow adoption of 3.5 until these are available. HP, IBM, Dell, etc. please get your SIM, Director, OpenManage, etc. updated ASAP as everyone will want to move to 3i once the dust settles.
- Backup: Hmmm. Here’s the only place that should cause some big growing pains. I’d expect the usual suspects will eventually support using their backup software remotely. But again, not out of gates. In fact, while the server manufacturers will take 3-6 months to get their management agents aligned with the 3i CIM model, my guess is that it will take more like 6-12 months for the major backup players to figure out the new backup paradigm. Your best bet is to try VCB or look at a major ESX backup player like Vizioncore vRangerPro or Phd Esxpress to support 3.5/3i if you want to move on this quickly. Since VCB should work in a VM without additional plumbing and support all storage types (see p. 29 here) it’s worth a look in 3.5 if you haven’t already checked it out, especially if you’re already on Enterprise ESX licenses.
So why move? What’s so compelling here?
- 32 MB footprint (yes, it takes only 32 MB of code to make the entire ESX server run)
- New hardware will ship with a dual bank flash drive for firmware like redundancy. Blow an upgrade and just switch back to the code you were running 5 minutes ago.
- Security. No agents. No services. No access without secure authentication.
- No more customizing installations for new releases.
- No more drive failures or complicated SAN booting setups.
So what’s the plan? I think many sites will adopt a two phase approach:
- Move to 3.5 using the standard install with COS for production. Poke & prod at 3i in the lab and migrate the problem areas forward.
- Once the issues are sorted, move to 3i. That’s assuming the hardware vendors understand that we want this for our existing servers and not just for some SKU that includes 32 GB of memory and 8 sporty cores with lots of zeros after the $ sign. We want to stuff a USB drive in an existing blade or discrete server and go with 3i.