July 18, 20205 yr So I've had this SSD cache drive in use for about two years. 480GB Kingston Sata SSD. Recently (a few months ago), I started noticing that after a few minutes of sustained writes, the write speed would slow to about 5 M/s and the server would be unresponsive until the data was finished being written to the SSD. Since unRaid seems to "read ahead" (not sure what its called), this means the server would be unresponsive for minutes, at least. So, I migrated all the data off the SSD cache drive and retested (moved the docker file and appdata to the array). No real improvement, same result. I stopped the array and removed the SSD cache drive and put in a 500GB rotational drive, 5400 RPM. I was able to copy about 300GB of data which took about 30 minutes with no unresponsive occurring. So it's clear it was the SSD causing the issue, but the SSD had no SMART errors so I don't understand the issue? Is that just how SSDs fail?
July 18, 20205 yr Community Expert Have you periodically been running Trim commands on the problem SSD to free up unused cells? If not then it is not unusual for performance to degrade over time. The Dynamix SSD Trim plugin can help with scheduling this. It could also be an age related factor as cells internal to the SSD are starting to fail.
July 18, 20205 yr Author I have not been running trim. Thought I read somewhere that it wasn't required on newer versions of unRaid. Any way I can check the life left on the SSD. SMART data seems fine.
Archived
This topic is now archived and is closed to further replies.