| « Nov |
|
Jan » |
| Mon | Tue | Wed | Thu | Fri | Sat | Sun |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 8 | 9 | 10 | 11 | 12 | 13 | 14 |
| 15 | 16 | 17 | 18 | 19 | 20 | 21 |
| 22 | 23 | 24 | 25 | 26 | 27 | 28 |
| 29 | 30 | 31 | ||||
Upcoming Events:
I just got an email from a colleague from EMEA. He showed me the picture below. The numbers alone for the uptime on the servers is huge (over 3 years). What’s even more interesting is the app – it’s ATM control software for over 100 banks and it’s in the production environment.
3 years is only a little shy of the record that I’ve seen previously of over 4 years of uptime.
If you have a host nearing the uptime record feel free to comment back here.
I try and write something new 1-2 times per week. I'll update more when I'm in the office and possibly less when I'm out and traveling. If there's something you'd really like me to hear my thoughts on then just email me.
Comments and FeedbackAnyone is free to comment. I will not filter or delete any comments unless they are obvious spam or flames. When commenting, any information you can share about who you are, how to contact you, what you do for a living, etc. would very much be appreciated. Let's keep the discussions live and interactive.
December 17th, 2008 at 7:55 am
wow, what about patching etc?
December 17th, 2008 at 8:00 am
mmmm tried to change the date of my test Virtual center server and got an uptime of 2560 days.
December 17th, 2008 at 8:47 am
While these stuff really impress me, it also scares me in the same time. A 3 years uptime of a host also means an unpatched system for all that time. Despite the security issues around that, my real concern would be: what if they ran into a problem to the point where they will need technical support from VMware? I believe they won’t get it, am i right?
I think I’ve seen a very similar screenshot for a bank in Saudi Arabia, but I don’t think it’s the same client.
December 17th, 2008 at 8:49 am
i added my reply before seeing Duncan's, so i assume we share the same concern
December 17th, 2008 at 9:09 am
Mike – really interesting. Out of interest – what's the host and guest technology – Windows / Linux etc?
December 17th, 2008 at 10:54 am
Mike – I believe the customer listed in your bigger picture openly acknowledges that they run unpatched, and how that isn't really a best practice. What these types of references DO is give VMware the ability to say “this stuff really is the most stable platform you can run.”
December 17th, 2008 at 1:13 pm
Yes, this case like the other ones I've shown before run the servers unpatched. They do this for a couple of reasons usually. The one that currently holds the record didn't patch because they run in a FDA (US Food and Drug Administration) regulated environment. Basically they have to produce a mountain of paperwork to certify a particular set of hardware and software configuration is valid for testing. When working with VMware they certify and product the paperwork for the ESX host once and then they have a certified VM environment. Since the VM never changes hardware they can just do 1 certification and then mass deploy that VM on any hardware as long as that hardware is also running the validated ESX host environment. If they patch the host though then they need to go through the paperwork trail and time and effort validating that new ESX host. This is why they never patched the systems. For this customer it was a calculated risk. The time saved in not patching and revalidating the hosts greatly outweighs the need to move the VM to an up-to-date ESX host if they need to for troubleshooting purposes.
To Hany's comment if the customer called in and didn't have an up-to-date system then VMware support usually does the best thy can to see if there's some glaring issue. If they think it might be a bug resolved in a current version they'll ask you to move it to a current system or patch your existing system. For these two customers they just cross that bridge when they come to it. If support asks to move it then they can just shut the VM down and move it or if the two hosts are VMotion compatible they can move it live. Again, it's all a calculated risk.
The bottom line to all of this is that ESX is stable enough that you can have it up and running with load for longer than you probably want to own that server anyways.
December 17th, 2008 at 1:14 pm
That's too funny. I guess you get the new record then.
December 17th, 2008 at 1:14 pm
The host was ESX. This was version 2.0.1 I believe in the picture.
December 17th, 2008 at 1:55 pm
WOW! That is amazing. I get the hosts updated frequently, most recently the ESX 3.5 series. So, I'm happy with my 75 days uptime right now!
December 17th, 2008 at 3:41 pm
Mike,
I remember a HIGH ranked VMware exec complaining about this once: “things like this mean these customers aren't buying the new features in the latest releases! This is not good!”.
I think he was joking. I think ………………………………………………………….
Impressive though.
Massimo.
December 17th, 2008 at 4:53 pm
Yeah, well, at least these two environments at these two customers aren't getting the new stuff. Of course VMware's competitors still don't have many of the features that ESX 2.0.1 had 4 years ago.
December 19th, 2008 at 12:59 am
I worked for the “1516 day” company up until March of this year
Patching and keeping hosts up to date is certainly the best practice but as you said in large environments things can turn slow for certain reasons.. we went from 2.01 to 2.5.1. to 2.5.4. Once we began down the 3.01/3.02 train we commenced patching with a bit more frequency. That particular host was rock solid (DL580G2) running 2.01.
December 19th, 2008 at 7:02 am
Thanks for the insight, Keith. It's always nice to have the customers commenting to validate things. It's a shame we can't get more case studies with customer names out there but at least we can anonymously share information this way.
December 19th, 2008 at 7:59 am
I worked for the “1516 day” company up until March of this year
Patching and keeping hosts up to date is certainly the best practice but as you said in large environments things can turn slow for certain reasons.. we went from 2.01 to 2.5.1. to 2.5.4. Once we began down the 3.01/3.02 train we commenced patching with a bit more frequency. That particular host was rock solid (DL580G2) running 2.01.
December 19th, 2008 at 2:02 pm
Thanks for the insight, Keith. It's always nice to have the customers commenting to validate things. It's a shame we can't get more case studies with customer names out there but at least we can anonymously share information this way.
January 2nd, 2009 at 7:28 pm
yo…
wonderful…