Brand New Dell PowerEdge T110, 4GB RAM
I have bought four Seagate Barracuda XT 2TB disks, and want to use the built in S100 Software based RAID controller.
The story so far:
-Enabled RAID in BIOS
-Created RAID5 partition of the four disks (5,6TB)
-Enabled UEFI boot manager
-Installed Windows 2008 R2 with driver for controller
-everything worked fine, the OS run fine on a C: boot partition
-Success, or so I thought....
After running the server for a few hours, copying files to it and doing some basic configuration -suddenly it froze. Had to press the button to restart it - not good...
On startup the RAID Bios indicates that the Virtual disk is degraded - Server apparently running on just three of the four disks in the Raid5
when I run Openmanage Server Administrator in Windows, I can see that the Virtual disk is Warning and the status is "Rebuilding". The first time, the server hanged again when it was about 75% complete of the rebuild. The rebuild process seems to take about 36 hours or so, not too impressive (is this normal?)
So my question is of course, what might be wrong? Are the Seagate disks not compatible with the S100? They are detected as they should, and everything is running perfectly well until the server freezes. Could it be related to the Read / Write cache that is selected when the volume is created?
Any hints and tips would be highly appreaciated
Best regards, Stefan
The only drives that are officially supported and compatible with the S100 controller are Dell OEM drives. The Dell OEM drives are different due to the fact that they have Dell firmware installed on them. Purchasing other drives "should" work but since you are having issues with the drives compatibility may be the problem.
Regarding the time the rebuild takes that is normal due to the size of the drives and the software RAID controller that is being used. Since it is software RAID it is dependent on the operating system for all tasks. The other factor that makes the rebuild time high is the size of the disks and the amount of data on the array, since the array is new and doesn’t have much data I would attribute the long rebuild time to the disk size.
The recomendation would be to puchase Dell OEM disks so you can ensure compatibility with the controller although the price will be much higher than the drives you purchased.
Thanks a lot for taking time to answer. The server is actually running quite well on the remaing three disks in the RAID-5 array during rebuild. I have now tried to replace the failed drive with another 2TB disk, working with the theory that there actually might be a problem with the drive itself, even if it is brand new. The rebuild is still in progress.
The errors that I have in the eventlog when using the other (defect?) drive are:
I have only seen errors that are related to disk 0, the other three disks are not generating any logged errors.
Regarding the array rebuild time, is there a way to set the priority for the rebuild process to speed things up? The server is idle so it could definitely use more CPU-power if it is possible to configure somehow?
You are very welcome. The errors that you are seeing in the SEL, specifically the Disk medium error likely mean there is a problem with the drive, but sometimes these types of problems can be corrected by updating the firmware on the drives if there is an update available.
Unfortunately with the S100 the option to adjust the rebuild rate isn’t available. With the higher end hardware based RAID controllers the option does exist and you can manually adjust the amount of the RAID controllers I/O that is used for a rebuild.