Here are the specs of my nas build:
Motherboard: ASRock J3455B-ITX (Bios version P1.20)
RAM: 8GB DDR3 Synchronous 1600MHz (0.6 ns) (2 x 4GB)
CPU: Celeron J3455 (Quadcore, 1.5GHz)
System Drive: Samsung SSD 830 (64GB)
Expansion Slot: DeLock PCIe x2 SATA controller (10 SATA Ports) (https://www.delock.de/produkte…kmale.html?setLanguage=en)
- 2 x 2TB SSD (SanDisk SDSSDA-2): RAID 1
- 3 x 5TB HDD (TOSHIBA HDWE150): RAID 5
- 3 x 8TB HDD (TOSHIBA HDWG180): RAID 5
OS: Debian 9.13
See full system-report attached.
Currently both RAID 5 RAIDs are degraded running with only 2 out of 3 disks:
When writing large amounts of data to one of the RAID 5 devices I experience a huge IO/wait load, load average goes up somewhere between 5 and 10 (see output of top below). Copying large files from one RAID 5 to the other RAID 5 is very slow - about 25 MB/s.
dmesg shows repeating errors:
I don't know if these errors are because of some bottleneck or if it is a hardware error.
Can someone point me to where the bottleneck most likely is? Is it
- the CPU (computing the parity data)
- the SATA controller card
- the HDDs themself
- something else?
Thanks for your advice,