Unusual problem with PE2950III with 2x Xeon 5450, 64GB RAM, PERC 6i

Thank you sir for your response.  I had the same thing in mind, and the PERC 6/i is going to be replaced by some of the H710s I have on hand to use until January when on a new budget.

I am not sure if a disk patrol is happening or what the problem is, but I find it odd that it occurs on the weekends only.  That is what befuddles me.  I agree it is completely unsupported and I know it is uncharted territory, and I am fortunate it is working at all.  The Perc 6/i driver is reporting the disks as unsupported but it just logs it and goes on about its business.  Windows Update is disabled  for now from even downloading updates, so it is not requesting a host reboot based on that setting, but according to the Event Viewer, it is actually rebooting gracefully.  I am not getting the typical "An unexpected Windows shutdown occurred" when the OS does boot, and it is always the same disk that is not found after the reboot, which is, ironically, a Dell certified 1TB SATA disk in slot 0.  After the reboot, the PERC 6/i does not see this disk, which is configured as the boot disk, so it hangs on the BIOS/POST/Pre-boot screen.  If the server is powered off and cold booted, I allow it to do the complete memory test and no errors are found, and when the PERC 6/i BIOS boots, it states that the configuration has changed and asks to press C to continue or F to import the foreign disk.  Before the power cycle, the "F" option does not appear, nor do any foreign disks appear in the RAID BIOS menus, nor the option to import them.

I appreciate your help knowing this is unsupported, and you didn't have to even assist, but you did anyway and I thank you.  If I can determine what it is rebooting (it doesn't seem to be a dirty shutdown) it may work until January when it can be replaced with a new server.  Meantime, I am going to replace the PERC 6/i with an H710 and Mini-SAS to SAS cables to the backplane.  It will be interesting if the 710 sees the 3GB disks as what the 6/i labeled them as (2TB) or if it will see them as 3TB.  If it does see them as 3TB, I am curious if it will allow me to extend the RAID0 array of the single disk to 3TB and it appear in Disk Management as a 3TB.  For anyone following this post, I will post my results.

Thank you again.

***EDIT - I missed your initial question.  No, there are no jobs that are scheduled for the weekend, or any interactive sessions, and this server has virtually no load on it.  I do think I will download the Event Logs and see if I can find a pattern.

Meantime, I am going to replace the PERC 6/i with an H710 and Mini-SAS to SAS cables to the backplane.  It will be interesting if the 710 sees the 3GB disks as what the 6/i labeled them as (2TB) or if it will see them as 3TB.  If it does see them as 3TB, I am curious if it will allow me to extend the RAID0 array of the single disk to 3TB and it appear in Disk Management as a 3TB.  For anyone following this post, I will post my results.

Yes, the PERC H710 supports HDs of greater 2 TB capacity and should be able to see all 3 TB.

Review the logs and let know of any unusual pattern

I did notice "Optimized drives" was scheduled weekly, so I suppose that could be crashing it.

Do you know what the largest Dell Certified SATA drive is for archival purposes (and not performance?)

And one last question and as I know you are busy.  Can the PERC 6/i be placed in HBA mode, disabling RAID functionality?  I think this was a function of the older cards, but I don't recall seeing it on the 6/i.

Thank you.

Will the H710 reject/hang the boot process with non-Dell drives?

Although the Windows GUI did not present the Unexpected Shutdown dialog (I anticipate this because the Shutdown Event Tracker was disabled on this server), the event log did store this:

The previous system shutdown at 6:46:48 AM on ‎11/‎26/‎2016 was unexpected.     EventLog 6008

The log immediately before this was: The Volume Shadow Copy service entered the stopped state. Event Source Name Service Control Manager    Event Source: Service Control Manager     EventID     7036  

There are several of these when the disk boots, probably as a result in 3TB drives on a 2TB controller.

I am seeing several of these Event 2335 from Server Administrator in the Windows Event Log:

Controller event log: PD 111(e0x00/s17) Path 1221000000000000  reset (Type 03):  Controller 0 (PERC 6/i Integrated) 

There are a few of these appearing after the server boots up.  Probably because of the same reason above.

So I was clearly wrong about it being a graceful shutdown as EventID 6008 shows.

I have enabled additional logging to show why this may be happening.  I don't see a particular error right before the shutdown occurs in the standard event logs or the Microsoft event logs.  I am starting to think it is one or a combination of 1) A predicted failure of the Dell Certified 1TB SATA Disk in 0:0:0 or the PERC is accessing a disk that it does not have the logic to access properly, causing a shutdown.  I could disable automatically reboot on STOP to see if that shows anything.  I believe I need to 1) Replace disk 1 that has a predicted failure, and 2) install the H710 and new battery.

You are probably not in a position to comment, but if you can, could you suggest a disk that may be most compatible with a Dell Certified Disk?  Obviously an Enterprise disk and not a Desktop drive.

Again, Thank you for your help.

I have enabled additional logging to show why this may be happening.  I don't see a particular error right before the shutdown occurs in the standard event logs or the Microsoft event logs.  I am starting to think it is one or a combination of 1) A predicted failure of the Dell Certified 1TB SATA Disk in 0:0:0 or the PERC is accessing a disk that it does not have the logic to access properly, causing a shutdown.  I could disable automatically reboot on STOP to see if that shows anything.  I believe I need to 1) Replace disk 1 that has a predicted failure, and 2) install the H710 and new battery.

Note that whereas the PERC H700/H710 controllers support hard drives of capacity greater than 2 TB, they are not supported controllers on the 2950 server. The desire to use 3 TB or more capacity hard drives will be accompanied by server upgrade to 11th Gen for PERC H700 or higher servers.

So far it has seemed to work without issues.  Only the replacement cables need to be used to connect to the backplane.  No unexpected errors like I am seeing here has happened, and so far it is running nicely.  I know the 2950III is well beyond it's supported lifespan, so I don't mind a little experiementing for lab environments, but in no circumstance would I place a server like this in a production environment..

Thank you

"Note that whereas the PERC H700/H710 controllers support hard drives of capacity greater than 2 TB, they are not supported controllers on the 2950 server. The desire to use 3 TB or more capacity hard drives will be accompanied by server upgrade to 11th Gen for PERC H700 or higher servers."

Are you referring to UEFI in this statement, or that the 2950III will just have a hard time understanding the disk size presented to it from the controller?

Thank you.

The PERC H700/H710 are not simply supported on the 2950 server regardless of the boot mode whether traditional BIOS or UEFI

Gotcha.  Thank you again!

The PERC H700/H710 are not simply supported on the 2950 server regardless of the boot mode whether traditional BIOS or UEFI

Note that the 2950 has NO UEFI support.