Debian Hardware :: HP P410 Controller On Wheezy (server Freeze)

I have a very weird problem with a HP DL180 G6 server that uses this controller.

The server has 4 SSDs configured as separate logical devices (no hardware RAID). I can install Wheezy just fine (creating a software RAID1 from sda and sdb). I installed zfs on it and installed a Windows 2008 VM on kvm with the disk image stored on zfs. The entire server froze when Windows was installing updates.

I rebooted and everything worked again, this time I launched iostat -x 2 on a separate ssh session, here's whet I saw after the server froze again:

Code: Select allavg-cpu:  %user   %nice %system %iowait  %steal   %idle
           0.25    0.00    0.19   20.83    0.00   78.74

Device:         rrqm/s   wrqm/s     r/s     w/s    rkB/s    wkB/s avgrq-sz avgqu-sz   await r_await w_await  svctm  %util
sda               0.00     0.00    0.00    0.00     0.00     0.00     0.00     7.00    0.00    0.00    0.00   0.00 100.00
sdb               0.00     0.00    0.00    0.00     0.00     0.00     0.00     7.00    0.00    0.00    0.00   0.00 100.00

[Code] ....

Thinking that zfs might be causing this, I destroyed the zpool, configured software raid on sdc and sdd, created a volume group and stored the image on a logical volume. The problem happened again.

I opened a ticket with the support of the company that provides the server and they updates the firmware of the controller. Now the server no longer froze while installing Windows updates, with lvm or zfs, so I went back to zfs.

Wanting to be sure, I launched iometer on the windows VM and set it to "4KB, 75% read 100% random" access pattern over 5GB of the Windows drive. The server froze after 7 hours.

I told the company that, and they replaced the SSDs with ones of a different model. I ran the test again and the server froze again, this time after 9 hours. I asked the company to replace the whole server. They did it (but kept the SSDs), I ran the test again and the server froze again, this time after 20 hours.

Now, this is starting to look like some kind of incompatibility between the HP controller and Debian rather than a problem with the controller. The settings are all default. It looks like the entire controller freezes because the system drives (sda, sdb) also show 100% busy despite not being under test from iometer.