Last week my unraid server had an unclean shutdown from a blackout.  Upon being able to restart the server, a parity check was started.  Before the parity check was completed there was a loud noise coming from the computer, but wasn't something I had heard before when a disk goes bad.  As I remember, I tried to stop the array (maybe having to stop the parity check first) to get the noise to stop.  The server has been on with the array operation saying "stopping" ever since.  Since then, I have ordered/received a new drive.  It is ready to install but will need to have a preclear done before using.  I assume because the disabled disk also says contents emulated that all the files are not necessarily lost.  Attached is a copy of the diagnostics.  I didn't see a smart report for the disabled disk, but am unsure of where to go from here.  Any help would be appreciated!

bluedragon-diagnostics-20190417-1716.zip

Disk2 appears to be failing, but then there was a problem with multiple disks dropping offline:

 

Apr 17 16:58:42 BLUEDRAGON kernel: md: disk1 read error, sector=9672
Apr 17 16:58:42 BLUEDRAGON kernel: md: disk3 read error, sector=9672
Apr 17 16:58:42 BLUEDRAGON kernel: md: disk4 read error, sector=9672
Apr 17 16:58:42 BLUEDRAGON kernel: md: disk0 read error, sector=9672

 

Reboot and post new diags.

I appreciate you taking the time to respond.  So, I clicked on reboot and confirmed.  Then looked at the cmd? side and it stated "The system is going down for reboot NOW!".  A few minutes later it said "php-fpm[1680]: [WARNING] [pool www] server reached max_children setting (20), consider raising it".  I don't know what this means and the server didn't reboot from what I could tell.  The array option on the main screen of the server webpage still says "stopping...".  Do I need to quickly hit the power button on the server computer to shut it down?  I believe I read on here somewhere that doing so will initiate a clean shutdown.  Don't know if this is true or not though so I haven't tried it.

Edited April 26, 20197 yr by Kinspappy

Type reboot on the console, if it doesn't after 5 or 10 minutes will need to force it.

Just typed in reboot on the console and will wait 5 - 10 minutes.

So, I went ahead and tapped on the server power button while on the console.  It said lots of things including "Running shutdown script /etc/rc.d/rc.0:".  A few lines later it said "Waiting up to 90 seconds for graceful shutdown...".  Then a few minutes later it said "Forcing shutdown...".  Followed by several lines with the last one being "Turning off swap".  Now it's been sitting like this for about 10 minutes, but it hasn't shutdown yet.  I took pictures of the text displayed on the console if it's needed.

I just tried the server webpage and it couldn't reach it, so I guess it's still trying to shutdown.

You'll need to force it.

Alright, I forced the reboot and uploaded a new diagnostics.

bluedragon-diagnostics-20190426-0404.zip

Parity and disks 1 and 2 are missing, a power cycle might bring them online, if not check cables, when they are back post new diags.

Yeah, I noticed they were not online and shutdown again to check connections and such.  Also installed the new drive that still needs a preclear.  OK, they are back online.  Here is a new diagnostics.

bluedragon-diagnostics-20190426-0440.zip

Disk2 looks fine, onboard controller is (at least partially) set to IDE, change to AHCI/SATA, then and since you already checked cables, and if the emulated disk2 is mounting correctly, rebuild on top of the old disk.

Good to know that Disk2 looks fine.  Just changed the sata settings under chip configuration from IDE to AHCI in the bios.  Rebooting now and will rebuild on top of the old disk2.  I guess now I have a new drive or 2nd parity drive once I preclear it.  Thanks for all your help!  If this works without any issues, I will mark this solved.

1 minute ago, Kinspappy said:

 I guess now I have a new drive or 2nd parity drive once I preclear it.  Thanks for all your help!  If this works without any issues, I will mark this solved.

It is worth pointing out that pre-clearing a disk is not necessary except as an initial stress test of a drive (although it is still a good idea if you are thinking of adding it as a new drive).    Rebuilding parity or a failed drive is in its own right a reasonable test of a drive so there may well be times when you want to forgo the pre-clear to get your array back in normal operation more rapidly. 

I understand.  I started the array and have access to all the files across all 4 disks, but disk2 still has a red X and says device is disabled, contents emulated.  I was about to do a read-check of all array disks, but now I am unsure of why it still says this.

5 minutes ago, Kinspappy said:

I started the array and have access to all the files across all 4 disks, but disk2 still has a red X and says device is disabled, contents emulated. 

That's expected, you need to rebuild on top, but before doing it check that the emulated disk mounted and contents look correct:

https://wiki.unraid.net/Troubleshooting#Re-enable_the_drive

I was following some newer instructions here that I didn't understand very well.

https://wiki.unraid.net/Replacing_a_Data_Drive

Either way, now it says parity sync/data rebuild started.  So, it looks like it's working.  Thanks again!  Looks like you have solved my problems!  Have a good night!