back to article Who actually needs virtualisation and collaboration tools

Email is now routinely referenced as a mission critical system. The importance of tools that help and encourage greater collaboration amongst staff, partners and even customers is also on the rise as organisations seek to raise productivity and performance. Is it possible for collaboration tools and e-mail systems to be utilised …


This topic is closed for new posts.
Gold badge

Virtual Tetris

Mail, SQL, MS LCS, WSUS, you name it. Configure once, and never again. Now, admittedly, we don't hammer our SQL servers enough for virtualization overhead to actually matter...(so the IOPS freaks who turn pale grey and then explode all over the comments section can get off here.)

We've been running all of our servers in a virtualized environment for about two years now. Exchange 2007 doesn't give us a spot of bother. Admittedly, it does unrepentantly nom all 8GB of RAM it's been assigned, and seems hungry for more. (I should really remember to up it's allocation one of these days.) It seems RAM hungry…but not especially disk I/O or CPU intensive. (Mind you, it’s running about 100 mailboxes, so this is an observation that is really only accurate for the SME space. Larger enterprises have vastly different loading.)

Our domain controller VMs are so small as to be irrelevant. They sit there, provide directory services, DHCP, and DNS. Live communications server is the same deal. No problems there. They are assigned 512MB of RAM and consume less than 256MB. I rarely see them eat more than 1% of a core. (What a waste a piece of metal would be for those VMs!)

Now…WSUS? This is a whole different animal. That is one hungry VM. It likes RAM. Lots and lots and lots of RAM. It also uses a more disk space than anything else I have operational. I have watched this VM pin two cores. Busy little beaver, that VM. This despite a couple or rebuilds.

SQL is just a database. The more you ask of it, the less you can share resources with other VMs. Virtualization is not an inherently negative platform for SQL…but you need to keep an eye on loading and responsiveness. (IOPS freaks…hush. Overhead is overhead; whether it is induced by the hypervisor or your SAN. I’ll accept a % or two of “inefficiency” in exchange for being allowed to be lazy.)

We’ve a few dozen “number crunching” VMs scattered hither and yon that run applications that are tragically single-threaded, and so 8 of them on a dual-quad-core system takes up less space than 8 systems. Huzzah. They don’t seem to nom much RAM, so if you play what we call “VM Tetris” just right you can often get up to 16 of these things in the right box. (Half of the VMs being ones that typically run their compute projects outside the hours of the other half.)

And that’s my story. It really all boils down to VM Tetris. As much as virtualization enables ease of administration and management, you still have to understand the workload of all your VMs. You have to understand that capabilities of your hardware. You pack your VMs in with other VMs in such a way that they won’t impinge on one another, and you can do remarkable things. As an SME sysadmin, I’ve not had the pleasure of using the really high-end management tools for long periods of time. I’ve no appreciation for mow much; (if at all) they take the “VM Tetris” part of the job off your hands. For a sysadmin like me however, (someone who can’t afford better than ESXi, “VM Tetris” is just another part of the job. Some days you get a box, or an L, some days you get a squiggly jaggy thing you have no idea where to put. And the damned lines only come when you’re not ready for them.


Errr... dont agree

dont know what the guy said above, its about 10x longer than the news items and looks 1000x times less interesting....

Anyhow, VMs are great. Any machine that does lightweight work , make a VM. You can reboot it and manage it without effecting the other machines. We have about 10 VM machine running off 2 servers. Because most machines work in burst, the physical machiens can comfortably provide the CPU / Disk IO / Network IO.

As for backups? Well, think the article was dumb. If you added a physical machine you would ensure that its protected against disaster. VMs are no harder. Its not a negative against the VMs the fact the can be spawned easily, thats a good thing, it just doesnt absolve your responsibilty to protect them.

Better stop now in case I write a book like the above....."virtual tetris my ass"

Gold badge

Wall Of Text

Early morning (as in before/during breakfast) comments lead to rambling. Brain has booted into the OS, but all the services haven't started. Like the service that edits posts, or the one that compacts gigantic walls of text into something smaller.

Those are dependant on the "second cup of coffee" service. Sorry to have bored you. Now to go make coffee...

Thumb Up


VMs are great, and people seem to think they are magically more reliable than physical systems. Possibly, sometimes, make a backup. The real point of virtualization is to be able to build any it infrastructure you want, on "a few, kinda powerful systems", without swapping cds, and finding missing cables.


glad you like vm's

but ranting about the poster above you and coming up with experience that most people have made 5 years ago, well, you text was even more useless than his.

but yeah, very much agreeable. servers need to be backuped.


Jobs Horns


wats this got to do with the price of fish?


the price? you can't handle the price!

Now that is a truly excellent question, the fishmongers of the world have been neglected for too long, it is an outrage, why, there should be an inquisition or something.

In terms of separation (the kevin bacon thing), I'd say it was like:

VM = Physical/Logical Abstraction Engine Software

Made/Sold by Companies

Wall St

Fish Price

In terms of "I'm an idiot, boil this down into something I can understand during the ad-break":

Think of two fish-mongers working at a pile of stinking fish, if one of them has a significantly smaller load of fish to work on, and they don't work together, he will eventually run out of work to do, the other will look on longingly as the one with less work to do goes to the pub. If we used a "supervisor" we can apportion the work between them efficiently and allow them to finish the load faster and everybody gets a pint!

The metaphor is: each fishmonger is a virtual machine with a hypervisor controlling access to resources. If we allocate them efficiently, we can all get to the pub sooner as the work is done with less bottlenecks! (overall) And we can save money, instead of hiring 50 fishmongers when there is really only work for 10.. (even if each fishmonger works on a specific type of fish and can't handle the others, in this case, each would be a virtual fish-monger running inside 10 real fishmongers operating 24/7 with no sleep.. )

Or, you could http://en.wikipedia.org/wiki/Virtual_machine before I start to make no sense.. (start, have I started yet?)



All these services need quite a lot of ram + IO

If you end up virtualizing services like that, re-run your numbers. 8/16GB+ lowend systems in a cluster will be providing much more power and run cheaper than buying the needed RAM upgrades in your host boxes (8GB Dimms ouch)

and NOONE stops you from having failover nodes virtualized etc. (let the cluster shutdown some dev boxes etc. etc.)

NPIV gives you all the added freedom you could need to P2V or failover into the VM side of things within minutes.

This topic is closed for new posts.