Looking at my RAIDs I see that the status bounces quite a lot from clean to active. There is a lot of activity on the RAID.
It’s a RAID 6 with 8 Hard Disks.
My guess is that is OK. From what the googles said, though a little murky, is that it is.
Is this happening because writing data causes the parity to be updated, thus becoming active, and after that it falls back to clean when done?
It’s just not crystal clear.
It’s not resyncing is it? I see no indications that it is.
Thank you all so much for your help!
With parity RAID, you’re most certainly going to get a lot of I/O because of parity calculations. Let’s also assume the drives themselves are not crapping out bad sectors left and right AND are “RAID-friendly” (TLER-capable/NON-shingled architecture/etc.). I would be suspicious of any thin provisioning mechanisms resident on the array especially if the base fs is say ZFS or BTRFS. Did you configure it with LVM first and then establish the file system?
Regardless, thin-provisioning + parity on spinning rust in this day and age is not ideal, IMO unless it’s being used for archival/dormant storage purposes. Just my 2c.
They are Enterprise drives made for a RAID. Anyhow, I looked at other RAIDs setups and they behave the same way and have been running for many years.
I was just curious if there was knowledge of major troubles if this was noticed.
Thanks for the reply. I’ll add your knowledge to my own and apply it in the future.
Thanks!
1 Like
I’m using LVM2 to provide RAID-10’s using NAS HDD’s and format them EXT4 and dump them into ProxMox as a folder. Then the ProxMox VM’s I deploy use qcow2 disk images and I format them inside the guest with filesystems that have data consistency algorithms (BTRFS, XFS, something like that). I don’t require a lot of I/O, just lots of capacity. This was the least “strenuous” design I could conceive being a Linux n00b and just flying by the seat of my pants lol! So far, so good, but it’s only been about 6 months. Time will tell. The only other thing I need to get to avoid a disaster is a UPS. Regardless of having low I/O demands and whatever, power outages have a funny way of making HDD heads skip sometimes because of like a microsurge riding in on the tip of the outage. The other thing is, when the power goes out, it quite often quickly cuts off, then comes on, then off, on off on. This really slams all electronics and appliances. Not a good thing to have a server plugged into lol