rebuilding mirror set with onboard RAID on S1200BTL

[idata]

Hi,

We have a server with an S1200BTL motherboard running Windows Server 2008. The system is configured with a mirrored (RAID 1) pair. One of the mirror drives failed. In the process of pulling the bad drive, the status of the good paid-half was changed to RAID 0. The system has been booted and is running with that single drive in RAID 0 mode. However, now we would like to reestablish the mirror paid with a new drive. The problem is that we can't get the system to let us change the status of the running drive to RAID 1, so we have no way to tell it to rebuild with the new drive.

We've tried using both the firmware setup (control-E during boot-up) and using the Intel RAID web interface. Neither one provides a way to change the RAID option on the running drive or to pair it up with the new drive. And the only option for the new drive is to also make it a Raid 0 drive (giving us two separate drives, not a mirrored pair).

I wondered about removing the RAID setup from the running drive, but I was afraid of losing data. The instructions aren't very clear about what happens when you remove a RAID set. I vaguely remember some training about this from 10 years ago, and I know there is a trick to making it work, but I can't find it.

Can anyone help?

Thanks,

Ben

[idata · Accepted solution]

Hello,

We regret to inform you that the procedure you are looking assistance for is not documented in our database neither in our documentation. Our recommendation is creating your back up first in case of data corruption and then recreate the RAID 1 array from scratch without initializing so the data remains on the array.

[idata]

That would be my main question: if I recreate the raid 1 array without initializing, should the data still be there? Of course I intend to do a complete back up, but it would help to know if I have a theoretical chance of success.

Thanks.

[idata]

Hello,

To answer your question, your affirmation is correct. Data will remain if you do not initialize unless we are dealing with data corruption already.