I'm trying to power down I am stuck. Problem started when I attempted to start a VM and GPU passed through which was ALSO being used for DOCKERs. (PLEX) Shame on me for not reading..but I am now stuck. It hung the VM service where it wouldn't load. I had a parity check finish so I wanted to reboot to get things running again. 

 

Stopping the array was my first idea, but no dice. It was stuck at 'stopping services'. I attempted to kill services via terminal, but that didn't help. Still stuck on that. I tried to do a shutdown from the GUI but that just lost me the GUI!

 

I still have SSH. I tried powerdown and reboot but nothing seems to happen. I believe the docker service may be running still? I also see the disk array appears to be mounted..so I'm clueless where to go. I will install open-files plugin next, but- any suggestions on what I can do NOW? Logs attached. 

zeus-diagnostics-20201208-0944.zip

Edited December 14, 20205 yr by dja

Lots of this in syslog:

Dec  7 19:56:11 ZEUS kernel: ata3.00: failed command: READ FPDMA QUEUED
Dec  7 19:56:11 ZEUS kernel: ata3.00: cmd 60/20:b0:b0:26:ef/00:00:01:01:00/40 tag 22 ncq dma 16384 in
Dec  7 19:56:11 ZEUS kernel:         res 40/00:58:90:a8:e5/00:00:cd:03:00/40 Emask 0x10 (ATA bus error)
Dec  7 19:56:11 ZEUS kernel: ata3.00: status: { DRDY }
Dec  7 19:56:11 ZEUS kernel: ata3: hard resetting link
Dec  7 19:56:17 ZEUS kernel: ata3: link is slow to respond, please be patient (ready=0)
Dec  7 19:56:21 ZEUS kernel: ata3: COMRESET failed (errno=-16)
Dec  7 19:56:21 ZEUS kernel: ata3: hard resetting link

Looks like a connection problem on disk4

Dec  6 10:18:22 ZEUS kernel: ata3.00: ATA-10: ST10000NM0086        00LF363      LENOVO,             ZA20CT6T, LB91, max UDMA/133
...
Dec  6 10:19:35 ZEUS kernel: md: import disk4: (sdd) ST10000NM0086_00LF363_LENOVO_ZA20CT6T size: 9766436812 

You should fix that after powering down. Might have to use the power button.

Hmm...That is a brand new disk with brand new cable. I can navigate to the drive without issue as well right now.  Not saying this isn't it- but the problem started once I started the VM with GPU passed into it. Everything else was working ok.

 

Is this really the only option? Parity rebuild has been close to 2 days and I just finished one. 

1 minute ago, dja said:

brand new disk with brand new cable

Bad connection doesn't mean anything is wrong with the disk or the cable. Could just need replugging. Check power connection too. Make sure there is no tension on the cables that might pull on the connection, etc.

Can I bring the array offline so I don't trigger another parity check? 

So, there is a way to shut down with no gui and everything else basically hosed....I went into terminal ON the server itself. (I don't know that it matters...) 

shutdown -t 5 now 

This worked and DID NOT require a parity check!!