Intel RST - Inserting an empty drive into the RAID1 removed data from the second drive

[IWoda]

Hello,

I have an issue with Intel RST on Windows Embedded Standard 2009.

I have a RAID1 array with two 500GB hard drives and one CF card as my system drive (only WES2009 and programs).

The RAID is used for data storage and as the system is offline (no outside network connection) and embedded the way to extract the data is to remove one of the drives, copy the data from it, format it and place it in the machine again.

Doing so (placing a formatted drive into the bay) instead of causing the data from the drive that stayed in the machine (which I considered to be the "master" disk) to be copied to the empty drive, removed all the data from the "master" disk.

Is this a configuration error or am I missing something about the concept of how Intel RST is supposed to work in RAID1 mode?

Would the behavior be different if the inserted empty drive was not the exact same one as was removed earlier?

Thanks in advance for all the answers.

[DArce]

What version of Intel(R) Rapid Storage Technology is installed and how did you trigger the RAID 1 rebuild? Was this inside the operating system or the the Control I Option ROM (before the operating system loads)?

[IWoda]

All the configuration was done inside the OS through the tray application. The rebuild was triggered automatically upon system start after reinserting the disk.

The About page for the OS app shows Intel RST 10.1.0.1008.

[DArce]

For the configuration and usage of this RAID 1 in your system, what you should actually do would be a Recovery volume which does actually use a Master/Recovery relationship between the drives.

In this case you should just need to remove the recovery drive to copy the data to another system and then put it back to the system and sync it with the Master drive.

While using simple RAID 1, and triggering a degrade state (removing one of the drives), may cause conflicts when the "removed" drive is placed back into the system. Even though it has been formatted, meta data on the drive can still be preserved and this drive can be confused as the drive with the more "updated" information and the rebuild can be done from an incorrect source.

For information on how to enable this, please check these articles:

http://www.intel.com/support/chipsets/imsm/sb/CS-026142.htm http://www.intel.com/support/chipsets/imsm/sb/CS-026142.htm

http://www.intel.com/support/chipsets/imsm/sb/CS-028485.htm Intel� Rapid Storage Technology (Intel� RST) — No option to create a recovery volume

[IWoda]

Thanks for the information.

I am aware of recovery mode and I have tried it as well, but I suppose during testing I got confused by the descriptions of 'continuous' and 'on request' modes.

What I want to achieve is a fully automatic resync of the drives upon inserting the new drive (a fresh drive or the recovery drive in case of recovery mode). This is necessary as the device has no user interface available.

From my testing even continuous mode required a manual triggering of updating the recovery drive with the data from the master drive upon insertion. Again, all configuration was done from the OS tray app. Is continuous mode supposed to always force the data from the master to be copied to the recovery volume or a new drive without the need to reconfigure the array?

Would inserting a fresh drive guarantee a correct resync in RAID1 mode?

[DArce]

From our testing if the drive was just removed from the RAID array (which causes a degrade state) and then placed back to the configuration, the rebuild process will start by itself with no user intervention.

Inserting a fresh drive, would require you to manually set the drive as an spare drive or rebuild to that new drive.