On Mon, Nov 09, 2009 at 02:35:53PM +0100, Sebastian Hagedorn wrote: > --On 9. November 2009 14:10:54 +0100 Simon Matter <simon.matter@xxxxxxxxx> > wrote: > >> While virtualization has advantages it has also disadvantages. One thing >> is that it introduces an additional layer of complexity into the game. >> It's my impression that in many areas virtualization gets introduced not >> because of technical reasons but because of political pressure. > > In our case I wouldn't necessarily call it political pressure ... it's more > like organizational pressure. We have fewer personnel resources than we > used to, and have to run more systems with them! > >> For a high power, mission critical system like a mail cluster I'd stick >> with real iron as long as possible. That may sound old fashioned but is >> what I would do after everything I've seen. You will need the irons >> anyway, with or without virtualization. Did I miss something? > > Maybe. Ideally you save irons by putting more than one VM on each. For the > mail cluster that may or may not be an option. I think it might, because as > I mentioned the current boxes are 5+ years old. So I'd think with brand new > hardware we would get away with less than 100% on each box. > > The main advantage that ESX would offer is in employing VMotion, VMMware HA > and such. It adds a layer of complexity, but also a layer of security and > convenience. > -- > .:.Sebastian Hagedorn - RZKR-R1 (Geb??ude 52), Zimmer 18.:. > .:.Regionales Rechenzentrum (RRZK).:. > .:.Universit??t zu K??ln / Cologne University - ??? +49-221-478-5587.:. I will go ahead and chime in here. If and only if your I/O usage is well understood and managed, would a VM option work. As others have mentioned, any amount of heavy I/O will take out or slow dramatically every VM that needs any I/O at all. Most people tend to over provision CPU/memory on VM host boxes. Usually phrases like "it's all in memory", "next to no I/O",... preceed the I/O backlog and consequent multi VM outage. People really like their mail to stay up, maybe slow but up. Regards, Ken ---- Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
