Single disk RAID 0 setup without data relocation

Mar 8, 2018
1
0
10
I have a number of systems with 20+ SATA drives (8-10TB in size). They have less than 30% free space available. The drives are behind a RAID controller with cache, but does not support JBOD cache modes. The data is already backed up and secure elsewhere, so any failure (even of RAID 0) is not a concern in this scenario. To enable the RAID cache for writes the drives need to be setup as single disk RAID 0's, which in essence acts like a JBOD w/ cache for writes.

There is a process that has been called 'fake init' where you can trick a controller into skipping initialization of a RAID 1 volume after creating it. Boot the server, insert a second disk, and the mirror takes over to repair the array, making it healthy. This isn't in the instructions obviously for the card, it's a 'hack' to get around starting with an empty array.

I'm looking to do something similar with RAID 0. The controllers in question are various, PERC, LSI, HPE SmartArray, Cisco. Assume in every case that JBOD cache mode is not supported so please leave out suggestions to use JBOD cache modes.

Is it possible to get any of these controllers to setup a single disk RAID 0 without wiping the data off the drives, even if it's a hackish workaround?

Thanks in advance.
 
Solution
A RAID 0 is data striped across multiple drives.
You can't extract a single drive and have the data usable.

Your procedure might work for a RAID 1, because the data actually exist, in toto, on each drive. You just have to fake out the controller.
A RAID 0, a particular "file" does not exist on only one physical drive.

USAFRet

Titan
Moderator
A RAID 0 is data striped across multiple drives.
You can't extract a single drive and have the data usable.

Your procedure might work for a RAID 1, because the data actually exist, in toto, on each drive. You just have to fake out the controller.
A RAID 0, a particular "file" does not exist on only one physical drive.
 
Solution
Also, if you're going to enable write caching, be sure you have a UPS connected to this. A power failure during cached writes leads to corruption and loss of the file being written best case, damage to the filesystem worst case. The situation has improved with journaling filesystems like NTFS and ext3, but I don't know how RAID 0 could affect that. Since the filesystem layer sits above the RAID layer, cached writes during a power failure could cause damage to the RAID array even if the filesystem on a single disk would've been able to survive it.