RocketRaid 2310 KVM

[Gerhard Hannemann]

Hi

Have an odd time getting Freenas to boot.
I build FreeNAS 8 as a KVM guest on a Xeon based host running on an Intel Q67SW board. It run well until I pass the rocketraid 2310 through to the guest. (This board support Vt-d).
This work on a normal debian guest and have experimented with ZFS on Linux - with success!
However FreeNAS just hang on what appears to be looking for the attached hard drive.
One clue; a few lines up in the boot messages, is say something like RR23xx no driver attached, and the message stop at hptrr channel [0 3[ start successfully - then is stops

Any ideas what is happening here. Is there are driver/firmware missing or not setup correctly

Regards,
Gerhard

 

[Gerhard Hannemann]

Hi

OK, judging by the response, it seems that nobody has and ideas of what is happening here.
This is what I found:
The Vt-d pass through was not correctly set. That is, the host reboot lost the setting required to pass-through the RocketRaid 2310 device.
Secondly the RocketRaid 2310 does not seem to be well supported - it was very annoying to deal with - this was probably my lack of understanding how this device can be use in a non Windows environment.
The JBOD function did not allow BSD environment to see individual drives - not useful. Using legacy drives only present other issues with this card.

In the end, it did not use it.
A PCI video card was installed as a second RocketRaid 2720, ( i happened to have), was installed to the PCI-E Grapics slot. (This cars is a PCIEx8 and this mother board only has one slot that can cater for this size). A little more fiddling, It finally works.
Freenas 8.3.x, 9.0 and NAS4Free all see this card and individual drive directly.
Now I have to find out why vtnet is not available on 8.3 - another post.

 

[cyberjock]

Not to sound like I (or anyone else) is ignoring you. But when it comes to virtualizing FreeNAS there seems to be a very well defined line. Either you know what you're doing and you can manage it on your own without help, or you don't know what you are doing and you stand a very high risk of losing your data. I realize you were trying to save yourself some troubleshooting if someone had the answer already, but mostly virtualization questions are ignored by the forum. Very few people have experience with virtualizing, even fewer do it in a fashion that I'd consider "safe" for their data, and even fewer feel like trying to answer questions because virtualizing adds so much complexity its quite difficult to provide support in a forum environment for it. I don't try to provide much support regarding virtualizing since I'm not too terribly knowledgeable and I've never used KVM.

You are also the first person I've seen in a very long time try to use KVM. Even our warning to would-be virtualizers mentions that KVM has had issues... http://forums.freenas.org/threads/a...ide-to-not-completely-losing-your-data.12714/ People had it work great for months, then suddenly their data is gone. No warning, no clue what went wrong. Just gone.

I've tested many Highpoint controllers, and none of them did vt-d well. All either failed to work right off the bat or worked, then suddenly stopped working until a cold boot. Often when the cold boot is completed your zpool is unmountable and horribly corrupted.

 

[Gerhard Hannemann]

Understood - thank you
This is more an experiment - I do have all files backed up - and shortly plan to get myself a dedicated NAS.
The problem is to get a decent (price and|or condition), M1505 base board here in Australia. The Highpoint decision was more base on that all that I had available at the time.
My old server is a standard Linux (Debian 5 -> 6) based software mirror raid and runs kvm. It has server web and databases servers for years without any issues. This new server required some fiddling to get going and that alone is already making be a little dubious as to it stability - well see how it goes.....
I already have a USB key to boot the NAS os from, just in case. I also discovered that the 2310 card would have caused a mess as parts of the hard drive are not usable if the disk are used on legacy mode.
Anyway - have read the above links. Hopefully I won't be another victim in the near future.