We recently retired a DB box at our data center that was in a Supermicro 70 enclosure with a 90 bay jbod attached.  This beast had 156 drives with a raw capacity of about 3/4 of a petabyte.  Previously it was just a bunch of zfs raid2z(RAID 6) pools of 4, 6 and 8TB devices as 24 different file systems.  I'd like to turn it into one massive storage bucket with a minio docker container running on it so it can act as a big S3/Glacier style object store.  Obviously with that many drives, a dual parity system is not enough redundancy.  What, if any would be the drawbacks of replacing the HBA cards with MegaRAID cards and doing hardware RAID 6's presenting themselves to unRAID as individual drives(assuming that is supported)?  

 

I do realize there are many other ways to do this without using unRAID, but just figured I would check to see if this was an option.

 

thanks,

 

-dev

Some of us have used old Areca controllers to allow us to use several old small drives and bundle them up and present unRaid with a single large drive. For example 4 old 2tb drives can be hardware raided together into one 8tb drive and used as parity. It adds a level of admin complexity, but it has been done.

Sent from my Nexus 6 using Tapatalk