| Forum Home | ||||
| Press F1 | ||||
| Thread ID: 111074 | 2010-07-12 23:30:00 | 1U Rackmount Mini-ITX Mobo + CPU that supports virtualization | Chilling_Silence (9) | Press F1 |
| Post ID | Timestamp | Content | User | ||
| 1118312 | 2010-07-13 06:51:00 | Any reason why? 1, its an established virtualisation platform with support 2, performance is better on non VT enabled cpu,s 3, greater hardware compatibility list 4, upgrade pathway to VMWare equivalent vMotion |
SolMiester (139) | ||
| 1118313 | 2010-07-13 07:26:00 | That motherboard looks like ATX though dpn't you want Mini-ITX? What about one of these (www.computerlounge.co.nz) Zotac boards? |
stormdragon (6013) | ||
| 1118314 | 2010-07-13 07:42:00 | Potentially even better! I'll have a nosey, thanks! | Chilling_Silence (9) | ||
| 1118315 | 2010-07-13 08:21:00 | 1, its an established virtualisation platform with support 2, performance is better on non VT enabled cpu,s 3, greater hardware compatibility list 4, upgrade pathway to VMWare equivalent vMotionHate to break it to you Sol, but: So are both OpenVZ and KVM (which Proxmox uses) OpenVZ beats the socks off Xen performance-wise, because it's not trying to virtualise the whole OS - it's basically chroot on steroids. OpenVZ also has greater hardware compatibility, because it doesn't need hardware virtualisation extensions at all. Upgrade path to any other jails-based solution, and also to Xen / VMWare with the addition of a kernel.You're right on points 2, half of 3, & 4 where KVM is concerned though. |
Erayd (23) | ||
| 1118316 | 2010-07-13 10:23:00 | Hate to break it to you Sol, but: So are both OpenVZ and KVM (which Proxmox uses) OpenVZ beats the socks off Xen performance-wise, because it's not trying to virtualise the whole OS - it's basically chroot on steroids. OpenVZ also has greater hardware compatibility, because it doesn't need hardware virtualisation extensions at all. Upgrade path to any other jails-based solution, and also to Xen / VMWare with the addition of a kernel.You're right on points 2, half of 3, & 4 where KVM is concerned though. Well thanks for the info Erayd, however support on open source is more about the users, of which Citrix Xenserver is considerably more widely used... Im not sure what you mean by virtualising the whole OS, xenserver is a hypervisor, nothing to do with any o/s? not sure where you are going with this?.....explain? jails-based?... |
SolMiester (139) | ||
| 1118317 | 2010-07-13 10:45:00 | ...however support on open source is more about the users, of which Citrix Xenserver is considerably more widely used...No doubt about that - Xen definitely has a larger user base. I thought you were talking about commercial support though, hence my comment. They both have readily-available commercial versions / support though - OpenVZ mainly in the form of Virtuozzo (www.parallels.com). Im not sure what you mean by virtualising the whole OS, xenserver is a hypervisor, nothing to do with any o/s?I mean that each guest is fully virtualised, including all the I/O, and each guest runs an entire operating system, including its own kernel. With OpenVZ that's not the case. not sure where you are going with this?.....explain? jails-based?...By 'jails-based', I mean any virtualisation technology that works on a process level rather than hosting an entire OS. OpenVZ is an example of such an approach - each 'container' (i.e. guest system) is completely isolated from all the others, but they all share a common kernel. The segregation is process-based rather than VM-based. It's basically chroot with proper resource management. This (en.wikipedia.org) page provides a pretty good overview of it. BSD jails is another implementation of the same concept. |
Erayd (23) | ||
| 1118318 | 2010-07-13 11:04:00 | No doubt about that - Xen definitely has a larger user base. I thought you were talking about commercial support though, hence my comment. They both have readily-available commercial versions / support though - OpenVZ mainly in the form of Virtuozzo (www.parallels.com). I mean that each guest is fully virtualised, including all the I/O, and each guest runs an entire operating system, including its own kernel. With OpenVZ that's not the case. By 'jails-based', I mean any virtualisation technology that works on a process level rather than hosting an entire OS. OpenVZ is an example of such an approach - each 'container' (i.e. guest system) is completely isolated from all the others, but they all share a common kernel. The segregation is process-based rather than VM-based. It's basically chroot with proper resource management. This (en.wikipedia.org) page provides a pretty good overview of it. BSD jails is another implementation of the same concept. Hi, yeah, I meant the free version, so support from a user base is better....as to the rest, it sounds more like App virtualisation rather than an isolated server o/s guest which mainstream commercial firms would run....while pretty cool, isnt virtualisation as of a server per sa, bit like running java apps or something....which would explain the performance......Xenserver is hate to say it ( I'm VMWare fan) the best type 1 hyper-visor on the market! |
SolMiester (139) | ||
| 1118319 | 2010-07-13 13:07:00 | Hi, yeah, I meant the free version, so support from a user base is better....Oh for sure - they're both pretty good in that respect, but Xen's user base is definitely larger. ...as to the rest, it sounds more like App virtualisation rather than an isolated server o/s guest which mainstream commercial firms would run....while pretty cool, isnt virtualisation as of a server per sa, bit like running java apps or something....which would explain the performance......Not quite - the guest containers run their own version of everything from init upwards. The only (notable) constraint is they must all share the same kernel. One of the nice side-effects of this approach is it's feasible to run tens or hundreds of guests on a single physical server, and it also allows 'overselling' of hardware - you can run more guests than the hardware is capable of supporting, provided the resulting aggregate load is still under that limit. Xenserver is hate to say it ( I'm VMWare fan) the best type 1 hyper-visor on the market!:thumbs: |
Erayd (23) | ||
| 1118320 | 2010-07-13 21:39:00 | Oh for sure - they're both pretty good in that respect, but Xen's user base is definitely larger. Not quite - the guest containers run their own version of everything from init upwards. The only (notable) constraint is they must all share the same kernel. One of the nice side-effects of this approach is it's feasible to run tens or hundreds of guests on a single physical server, and it also allows 'overselling' of hardware - you can run more guests than the hardware is capable of supporting, provided the resulting aggregate load is still under that limit. :thumbs: Sounds great, but if anything happens to that kernel, you lose everything....not sure I would want that in a commercial environment. |
SolMiester (139) | ||
| 1118321 | 2010-07-13 23:21:00 | Sounds great, but if anything happens to that kernel, you lose everything....not sure I would want that in a commercial environment.Huh? Not quite sure what you're getting at there, can you expand on that a bit? In my experience, things don't just 'happen' to kernels, they are caused, usually by people who don't consider all the ramifications of messing with them. Doesn't your comment also apply to Xen? What if something 'happens' to the hypervisor kernel? |
Erayd (23) | ||
| 1 2 3 | |||||