[solved] 6.9 Rc2 updated, can't log into gui during parity sync

[greyday]

So I had a drive and a parity drive go down a couple days ago; after rebuilding the drive using the second parity and testing it all out, I decided to update to rc2 before adding the first parity back in. Everything seemed to go smoothly, the array looked fine after reboot (though I did have to start the array to then get to the point where I could enter in my key to actually start the array, which was a little weird) and after testing out the update a little, I went ahead and began rebuilding parity last night.

This afternoon I got a text (I have warnings set up to go to my phone) that the second parity drive and one of the main drives both had read errors. Not a huge number, seemed like a reasonable amount for a parity rebuild (less than a couple hundred) so I monitored it for a bit and since it wasn't throwing further errors I decided to wait and see if more occurred before halting the sync.

I just went to check on it and I can't even access the login screen, it times out. It showed up on a lanscan so I tried pinging it and aside from longer latency than usual it responded, so I ssh'ed in. Checking TOP it shows Unraid is using roughly 25% of the cpu (which is higher than usual but not insane, though I DO have a 3950x so that is a lot). Next on the list is mdrecover+ at around 6% cpu, I assume that's the parity sync? That's where I am at with it right now as I am not sure what other tests I can run or what I should do, other than shutting down which I would prefer not to do in the middle of a parity check if it's potentially still running.

I do have the 6.8.3 thumb backup and all essential data is also backed up, so there's no huge risk other than time and discs, but I'd like to know what to attempt before doing anything that nukes the whole system (and I'd like to reseat the DAS cables and card once I know it's safe to power down). The PiHole docker is still active and accessible, so the system is at least partially up.

 

I was able to do a diagnostics dump; there are some powerout notes from the APS (reasonable, winter storms and all) but none were long enough to trigger a shutdown, the whole server only uses about 20% battery capacity so there's a good 20 minutes of reserve. There are some weird "device reset" notifications that are concerning, though.

 

Physical set up notes: I have the drives installed in a Dell Xyratex HB-1235 12-Bay DAS connected via a SFF-8088 cable to LSI 9200-8e flashed for HBA (purchased from ArtoftheServer and been working fine for months).

 

Advice? I'd like to try any and everything before shutting down as I don't want to screw up the parity drives (not that I'm sure that would happen but I'd rather not find out...).

EDIT: I decided to just go under the assumption it was some weirdness from all the brownouts this weekend and rebooted, and everything seems ok so far. Parity sync has begun again and the system seems like it might be stable now; I'll come back and mark it solved if it makes it through the parity check/resync...

 

EDIT2: looks like it was an 8088 cable needing reseating. I'm gonna mark this as solved but if it repeats I'll come back.

Edited February 18, 2021 by greyday