Hyper-V Server 2012 USB boot best practices
- Thread starter BackWoodsTech
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Ok, note to those with windows 8 phones...initial post message body input geeks out on this forum.
Reply works fine. Anyway, I've read the instructions from MS regarding USB boot, and have actually done it, so this really isn't about the mechanics of creating a bootable USB or anything.
I'm in the process of building a new server. I'd say Hyper-V lab, but its not multi node, so its not a true lab. I'm the long run in hoping this will function somewhat as a SAN device that future computing nodes will boot VMs from. Ok, sorry for all of the aside.
My general concern with USB boot is the single point of failure aspect, and that leaves me inclined to run two small SSDs mirrored as boot. But that is a waste, they hypervisor just loads everything to memory. Its just the lack of redundancy that leaves me feeling its a little weak.
How about a second USB drive that gets an image of the boot USB flashed on a regular basis programmatically? Any thoughts about this issue or perspective from those in the industry on how they view this is greatly appreciated.
FYI, I'm just a lowly GIS manager, but I keep myself educated on tech in general in an effort to stay ahead of the curve and to bring any relevant implementations to the sever environment I manage. Hyper-V just happens to be something I've been having a lot of fun for the last year or so, but for larger, future deployments it is becoming much more relevant
Reply works fine. Anyway, I've read the instructions from MS regarding USB boot, and have actually done it, so this really isn't about the mechanics of creating a bootable USB or anything.
I'm in the process of building a new server. I'd say Hyper-V lab, but its not multi node, so its not a true lab. I'm the long run in hoping this will function somewhat as a SAN device that future computing nodes will boot VMs from. Ok, sorry for all of the aside.
My general concern with USB boot is the single point of failure aspect, and that leaves me inclined to run two small SSDs mirrored as boot. But that is a waste, they hypervisor just loads everything to memory. Its just the lack of redundancy that leaves me feeling its a little weak.
How about a second USB drive that gets an image of the boot USB flashed on a regular basis programmatically? Any thoughts about this issue or perspective from those in the industry on how they view this is greatly appreciated.
FYI, I'm just a lowly GIS manager, but I keep myself educated on tech in general in an effort to stay ahead of the curve and to bring any relevant implementations to the sever environment I manage. Hyper-V just happens to be something I've been having a lot of fun for the last year or so, but for larger, future deployments it is becoming much more relevant
Last edited:
See this thread - http://forums.servethehome.com/wind...tion/1062-running-hyper-v-2012-usb-flash.html
The last post has some info.
I haven't tried any further to get it to work, personally.
Some newer Dell servers support a RAID-1 type configuration of their internal camera card SD type storage. I know those work with ESXi.
The last post has some info.
I haven't tried any further to get it to work, personally.
Some newer Dell servers support a RAID-1 type configuration of their internal camera card SD type storage. I know those work with ESXi.
Jac, I am aware of boards like that, but it wont be anything I will have.
That thread really doesn't provide anything conclusive about best practice however. Again, I'm not looking for a 'how to do', this is about addressing the single point of failure.
Maybe I'll put on the flame suit and head over to Server Fault.
That thread really doesn't provide anything conclusive about best practice however. Again, I'm not looking for a 'how to do', this is about addressing the single point of failure.
Maybe I'll put on the flame suit and head over to Server Fault.
How about this?
Or if you have a spare 2.5" drive space you could use something like this:
Addonics Product: Dual CF - SATA HDD Adapter
I'm not qualified to speak to "best practices" but your concern about a single point of failure should be addressed by either of the above approaches.
If you had an unused USB header on your MB you could easily run two USB keys off of something like this (after removing the back panel plate):
Or if you have a spare 2.5" drive space you could use something like this:
Addonics Product: Dual CF - SATA HDD Adapter
I'm not qualified to speak to "best practices" but your concern about a single point of failure should be addressed by either of the above approaches.
I think Hyper-V USB Boot Best Practice is: don't do it (except for testing).
Hyper-V Server is basically windows-server core with most of its "roles" disabled. Windows server is not like ESXi (which generally doesn't write to its boot media after starting except to update config files) and Linix "Live" boots of KVM (which copy themselves to a ramdisk before booting). The file management and virtual memory management functions of Windows are mostly one in the same thing. Even if you are not "swapping" to disk due to memory over commit you will get lots of virtual memory activity in Windows due to general file activity and the way Windows dynamic libraries (DLLs) work.
Unless you are using a really fast USB3.0 memory stick on a good USB3 port your performance will suck rocks. You'll give the USB stick a premature death due to too many write cycles. And you'll be disappointed.
Testing OK. But as a "best practice" you really shouldn't run it that way. Go buy a small used Intel SSD and be happy.
Hyper-V Server is basically windows-server core with most of its "roles" disabled. Windows server is not like ESXi (which generally doesn't write to its boot media after starting except to update config files) and Linix "Live" boots of KVM (which copy themselves to a ramdisk before booting). The file management and virtual memory management functions of Windows are mostly one in the same thing. Even if you are not "swapping" to disk due to memory over commit you will get lots of virtual memory activity in Windows due to general file activity and the way Windows dynamic libraries (DLLs) work.
Unless you are using a really fast USB3.0 memory stick on a good USB3 port your performance will suck rocks. You'll give the USB stick a premature death due to too many write cycles. And you'll be disappointed.
Testing OK. But as a "best practice" you really shouldn't run it that way. Go buy a small used Intel SSD and be happy.
