Samsung 8TB QVO in raid?

[ycp]

Hello,
Anyone have experience with Samsung 8TB QVO drives in Raid configurations like Raid 0 or Raid 5?
Do they perform normally or are there any potential issues that can prop up?

 

[MBastian]

Apart from the QLC caveats and possible unresolved issues with the firmware(did you check?) I do not see a problem. Such an array would probably a slouch on I/O (write) intensive tasks but as a data dump it should be fine. I am questioning the RAID choice, especially Raid0 for obvious reasons. Apart from a redundand boot drive setup you should consider more modern options like ZFS.

EDIT: It seems many people are using them for their data hoarding needs without issues.

 

[ycp]

I would be using the RAID 0 for some non-critical workloads with nightly backups to regular hard drives.
Just looking for a cheap RAID 0 setup.
I have not checked about the firmware issues. I didn't know there were any.
Basically my workload is read heavy. Just fill up the array once and probably not write much to it afterwards

 

[Allan74]

Basically my workload is read heavy. Just fill up the array once and probably not write much to it afterwards

Glad I found this thread....and don't mean to highjack it....sorry.

Samsung 4TB QVO's are on for $170 USD locally right now and I am thinking about deploying a new (old) 2U SFF 24bay R720 or similar and starting with either 8 or more drives, to start moving my Plex library from spinning rust.

Hey ycp, how many drives are you considering using in RAID 5 ?

How would the average guy set something like this up ?
3x 8 drive RAID 5 ?
2x 12 drive RAID 6 ?
...something in between ?

thanks in advance,
Allan

 

[T_Minus]

I would never RAID0 any SAS or SATA SSD it makes no sense at all... just keep the drives setup as a single drive.
If you want performance use a NVME drive. If you want more performance use an Optane. If you need even more then get into DIMMS.

Unlikely anyone here is going to suggest RAID5 for anything, especially not 8 QVO consumer SSD... way too risky.