S5000VSA(SASR) Server board with Embeded RAID5 using W2k8r2

[JHov1]

We have installed a S5000VSASASR (S5000VSA) server board with 16 GB of ram, 4 Seagate Cheetah 15K SAS drives, setup as RAID5 using the embeded software RAID with activation key. The OS is Windows 2008r2. The problem is with the RAID configuration. It randomly corrupts and virtual drive has to be rebuilt. We have the latest bios (.98), latest firmware (0.46), latest backblane (2.14) and the latest W2K8r2 drivers installed (Chipset, LAN, RAID II). We even swapped out hard drives to see if anything changes and it doesn't. Sometimes, if you turn off the server and bring it back up, the RAID bios loses all information on RAID configuration and everything has to be virtually re-built. Needless to say, this is completely unacceptable and inadequate. I've spoken to Intel support. So far, they've replaced server board and RAID backplane but the results are the same. For a while, we thought the problem was being caused by the virtual server (Hyper-V) in Windows 2008r2 but we removed virtual server with no change in problem. Only one tech @ Intel semi-acknoledged the problem and admited to them knowing about this. Something about the LAN driver and the processor not being in sync (time-out) and hence, causing a RAID corruption. I've yet to find out any information on it nor have I been contacted by Intel with a solution. To say that we and the end-user who purchased the server are angry, is an understatement. concrete HELP would be greatly appreciated.

[JHov1]

Just tried it again. Clean Windows 2008r2 install, updated drivers and let the server sit over night. When I booted-up the server, one of the raid drives was in "ready" mode as if it was never virtualized and the hot spare drive was in "fail" mode. Why and how could a virtual drive of a RAID fail by being off for 5 hrs.?!?!

[Daniel_O_Intel]

That doesn't make much sense. I don't see why a NIC driver would have anything to do with the RAID, either.

Which hard drives are you using? The only time I've seen this is when I had mismatched firmware on some of my drives, even though they were the same manufacturer / model.

[JHov1]

I agree, it doesn't make any sense but it's still happening. The 4 drives installed are all Seagate Cheetah's model # ST3300657SS with firmware 006. I checked every one and they all have the identical build and firmware with no newer firmware available. I spoke to a tech at Seagate and he suggested I jumper the back of the drives to bring the speed down to 1.5. I did that yesterday so we'll see if that helps. It defeats the purpose of having fast drives but at this point, I'll take stable vs. fast.

[Daniel_O_Intel]

That's a good point. I don't think the VSA or the backplane support the SAS 6-Gb yet.

[JHov1]

Surely, we can't be the only ones having this ongoing horrific problem. I'm convinced this is a design flaw as we've exhausted every possible option suggested by Intel support. It has something to do with the on-board RAID-II, using SAS drives and Windows 2008r2. In Windows 2003r2, no problem. In non-sas, using SATA drives, utilizing the onboard RAID-1, no problems. The moment the SAS backplane in installed, RAID-5 is configured and Windows 2008r2 is installed, the problem starts. Sometimes, it takes it a few hours and sometimes a few days but it always ends-up in a RAID degration and drives start going off-line. Our latest attempt at fixing this problem was to get rid of the Seagate Cheetah 15k harddrives (which the exact model were apparently not on the certified drive list) and we switched to Fujitsu MBA3300RC 300 GB SAS drives and downgraded the backplane to firmware 2.12. At first, everything seemed fine and system ran great for about a couple of days. For a test, we shut down the server and left it off over night. When the server was brought back up, one of four of the RAID drives was offline and the Hotspare drive was marked as "unconfigured good". Windows 2008r2 crashed and we had to do yet another restore.

This is beyond absurd and is completely unprofessional by Intel. We've gotten so many different suggestions from Intel support staff and only once did someone actually acknowledge the problem as a known issue but offered very little in the way of a solution. This whole server setup is going to get trashed soon and Intel will be hearing from our attorneys on this matter. A class action law suit might be in order. We have wasted over 80 man hours of rebuild time, since the server was first installed in March of 2010. In addition, with the server constantly going down and it being the primary AD server for a business, end-user has lost countess number of hours and dollars.