Noticed logs filling up this morning so started to investigate and looks like i'm getting errors on one of my cache ssd's for at least the last couple of days.

 

I have pool of 2*1TB Sata SSD's, the one erroring is a bit older but only couple years i believe. I did flush edit:scrub the drive and continued to get errors it said were uncorrectable. is this a bad/dead drive? what can i do / check next to verify or try to repair this? any other suggestions for next steps?

 

My other drive in the pool is ~6 months old so i'm not too worried on using that by itself while i wait for a replacement if i have to. diagnostics attached

rlyeh-diagnostics-20230427-0737.zip

Edited April 27, 20233 yr by Azreal

Syslog rotated, reboot and post new diags after array start.

Thank you for jumping in quickly. See updated diagnostic.

 

Also worth mentioning - one of my array drives shows as missing. I don't think this is related but maybe it is. It was fine until i turned the array off prior to making the earlier post, then it said it wasn't detected and contents emulated. I have had some issues iwth this happening time to time and changing / reseating cables had helped stop that until today. Maybe relevant to the SSD issue maybe not. I had been planning on pulling that drive and replacing with a larger one but just hadn't gotten to it, the drive is empty anyways. 

rlyeh-diagnostics-20230427-0905.zip

Run a correcting scrub on the pool and make sure all errors are correctable, since it's an MX500 suggest updating the firmware, since it's been reported to help with these SSDs dropping.

thank you for the link. looks like the scrub fixed it this time, it did not earlier. not sure i understand why, reboot helped there?

 

firmware update looks like it will be fun. i have a windows pc i can put this in to do the udpates but man what a hassle. not seeing any way to do this with it hooked into the unraid machine but please let me know if that's wrong and there is a way. i'm probably mediocre in my linux skills and understandings at best so leveraging windows utilities probably is the path of least resistance here. 

 

thanks again for the assistance today. 

for anyone else that finds this - firmware update took ~1 minute on my old very dated win 10 pc - maybe not even that long. it didn't make me reformat or anything. threw me off at first because it wasn't appearing in my computer next to the boot drive but them remembered it's not formatted in windows compatible format. when i looked at the utility from crucial it saw it just fine. automatically identified the firmware is out of date and prompted to update it. 

 

one thing i probably could have avoided was starting the array after i removed the drive from the cache pool. lessons learned, pretty sure that was unnecessary but i was thinking it would need to rebuild anyways.... biggest headache was just pulling drives and bouncing between machines. everything else appears to have performed as expected.