Cache Pool goes Read-Only every few days

[cleight]

Good morning Community,

 

I am hoping someone smarter than me can figure out what is going on with my server. Every few days my cache pool goes into a read-only file system. I reboot the server and then things are fine again for a while. Now this morning this happened again and I rebooted the server, however this time it came back as Read-Only. I see no SMART errors on the 3 drives, and when I ran a scrub last week it found 6 errors but couldn't repair them automatically. I believe the issue might be happening when a TRIM operation occurs but not 100% sure on that.

 

I have attached two different diagnostic files of when this has happened, once from last week and one from today. As one can see from the dates it was both on Thursday's.

unraid-diagnostics-20190523-1310.zip unraid-diagnostics-20190516-1059.zip

[JorgeB]

There are several segfault and btfs checksum errors, first suspect would be bad RAM, run memtest, also make sure respecting the max RAM speed depending on the current config as some Ryzen systems are know to corrupt data with overclocked RAM.

 

[cleight]

2 minutes ago, johnnie.black said:

There are several segfault and btfs checksum errors, first suspect would be bad RAM, run memtest, also make sure respecting the max RAM speed depending on the current config as some Ryzen systems are know to corrupt data with overclocked RAM.

 

@johnnie.black Thank you for the quick reply. I will run a memtest this evening and see if it reports anything. As far as overclocked memory, that isn't something I setup in the Bios so I don't believe that would be a thing? Here is what shows in System Profiler: DIMM DDR4 Synchronous Unbuffered (Unregistered) 2133 MHz

[cleight]

@johnnie.black I let the server run memtest overnight it completed 2 full passes with 0 errors detected. Rebooted the server this morning and my cache pool is mounted in a writable state at the time. Any other thoughts/suggestions to this problem?

[JorgeB]

Since current pool has uncorrectable errors, backup cache, re-format and restore the data, if it goes read-only again after a few days very likely there are still hardware problems.