I have on e Beelink Me Mini Zima OS 1.5.1 installed.
There are four WD Blue SN5000 1TB installed with a RAID5.
If i boot the system the four drives are “healthy”. i copy 500-1GB and still all four drives are healthy.
Today i copied additionally 2 GB Data to the storage and then the transfer broke and two of four nvmes are shown disk broken. If i restart the system all four drives are up again (and all transfered data until the failure). I copied than the last failed file and it worked. Then again i copy data and after around 1-2GB data both disks going offline. After the reboot the disks are back and the data are there. Look for me for an SW failure.
Any tips?
I think this isn’t a ZimaOS software issue, it looks like the NVMe drives are dropping out under load.
Two of the four NVMEs only fail after 1–2GB of sustained writes, then come back immediately after a reboot. In my opinion, that behaviour is classic thermal or power-rail cutoff on small mini-PCs.
When RAID5 starts pushing continuous writes, the NVMe controllers heat up fast or the system hits a power spike. Once the controller crosses its thermal or power limit, the drive disappears briefly, mdadm marks it “faulty,” and after a reboot everything is cool again so the drives look healthy.
If this was a ZimaOS bug, all four drives would fail together or the RAID layer would show consistent errors, not two drives dropping only under sustained load.
You can quickly check it by running:
smartctl -a /dev/nvme0n1 | grep -i temp
If you see temps climbing into the 70–80°C range, that would confirm it.
In my opinion, this Beelink just can’t keep all four SN5000s cool or powered during heavy RAID5 writes. Better thermal contact, proper heatsinks, or limiting RAID write speed can help stabilise it.