Unraid crashes 10 seconds after booting up, with the attached screens coming up then the system is unresponsive.

 

I have replaced the RAM and run memtest to make sure it is not the RAM.

 

Unraid crashes 10 seconds after booting up, with the attached screens coming up then the system is unresponsive.

 

I have replaced the RAM and run memtest to make sure it is not the RAM.

Since it references BTRFS, I'd physically disconnect any BTRFS formatted drives and try again. Obviously that's not a solution, but it would be my first troubleshooting step.

Thanks for looking into it for me, unfortunately no drives are formatted BTRFS, I have never tried that file system.

Thanks for looking into it for me, unfortunately no drives are formatted BTRFS, I have never tried that file system.

Well, try disconnecting the drive that holds your docker image and see what happens.

I unplugged all the hard drives, and unraid boots and does not crash!

 

I will now start adding drives back in one at a time until I find the offending drive and replace it.  Hopefully that process will fix the issue.

I unplugged all the hard drives, and unraid boots and does not crash!

 

I will now start adding drives back in one at a time until I find the offending drive and replace it.  Hopefully that process will fix the issue.

Replacing the offending drive will likely not solve it, if it's one of the array drives. Still, it will be helpful to see what happens with various combinations.

All of the drives in this system are array drives, there are no cache drives or any other non-array drives.

 

I will report back my testing progress soon.

I have narrowed it down to one hard drive, and can consistently boot and stay stable overnight with the one hard drive disconnected from the motherboard.

 

However, when I try to start the array from the webGUI the system crashes again.  I would have thought with the hard drive disconnected it should be able to start the array.

 

What do I test next?

 

*edit*

 

And I already tried swapping sata ports and sata cables, no change still crashes if plugged in.

 

And when I tried to boot the array with the hard drive unplugged it unassigned the disk. 

 

I am trying different combinations of hard drives plugged in, but now have a problem that disk 8, the drive is unassigned.

 

How do I re-assign it without doing a parity rebuild?

 

 

*edit* 

 

It seems like the system will boot and be stable with any single hard drive unplugged, but crash every time they are all plugged in. 

Trouble is now, I have an unassigned disk and to get to the webGUI I have to have a drive unplugged.

How the heck do I get the hard drive re-assigned now?

Attach the diagnostics zip from the current situation. Also, try starting the array in maintenance mode and see if that changes things.

just throwing this out there have to tried a working hard drive in the sata connector of the offending drive?

 

it could be the the actual port not the drive.

 

It seems like the system will boot and be stable with any single hard drive unplugged, but crash every time they are all plugged in. 

Trouble is now, I have an unassigned disk and to get to the webGUI I have to have a drive unplugged.

How the heck do I get the hard drive re-assigned now?

 

One thing to consider is your Power Supply.  I am saying this because you indicated that if you unplug ANY drive the system will boot and stay up.  Apparently, you have 23 (or 24) drives on your system and that represents a heavy current draw when all of the drives are spun up when the array starts.  I notice that your PS is a 750W unit but does it have a single 12V rail?

I haven't really posted my suspicions here, I was trying to get more info before putting out my theory.

 

Based on what's been said so far, I think there is a corrupt docker image causing the crash. They are all array disks, so trying to start the array will cause the crash, no matter which disk is removed. If all the disks are present, the array tries to auto start. Removing one disk, any disk, causes the array to stay stopped until manually started. I'm hoping maintenance mode will allow the array to start without crashing. Also, diagnostics would be very helpful.

I will post diagnostics later today, thank you!

 

I have tried multiple ports and cables, no affect.  I also added a sata card and removed all the motherboard sata connections, same crashing effect.

 

This is my second system, 12 drives on a single 12v rail Corsair AX750 so lots of power.

 

So if this is a corrupt docker image (this server is running Plex docker), then booting in safe mode would deactivate this? 

I have already tried safe mode and it still results in a crash, but I will do some more tests with different sata connectors and cables and such with safe mode to see what happens.

 

 

I tried booting up in safe mode with all drives connected, crashes.

Booted up with my now-unassigned drive disconnected (from when I tried starting the array with a single drive disconnected, it unassigned the drive), same crash

Booted up with the unassigned drive and another drive disconnected, boots fine.

 

This is consistent behavior regardless of safe mode or not.

 

I have attached diagnostics from my boot in safe mode with 2 drives disconnected, the unassigned one, and another random drive.

 

*edit*

 

If I connect my SATA cables from both of the disconnected drives while the server is running, they are detected and no crash. 

When I assign the drive that was unassigned during my array-start attempt, it is "blue" in the menu and I get the option to start the array and rebuild the drive.

Any way to get unraid to trust the drive/array and get back to normal without a rebuild?  I'm not confident the system is going to survive a rebuild but would like to get the drive back assigned so at least that is back to normal?

 

 

 

Safe_Mode_Diagnostics.zip

Regular_Boot_Diagnostics.zip

So if this is a corrupt docker image (this server is running Plex docker), then booting in safe mode would deactivate this?

No, but this

try starting the array in maintenance mode and see if that changes things.

would.

I kept the un-assigned disk as un-assigned and clicked maintenance mode and started the array.

 

Started successfully and remained stable.

 

Next step?  Can I get unraid to trust the drive that is unassigned or do I try to assign it then boot in maintenance mode and rebuild?? 

 

Or something different?

 

 

Do you know which drive has your docker image on it? Does that drive have other data besides the docker image?

No, I cannot access the Docker section of the web browser GUI to determine which drive has the docker installed on it.

 

And yes that drive does have other data on it.

No, I cannot access the Docker section of the web browser GUI to determine which drive has the docker installed on it.

 

And yes that drive does have other data on it.

Open \\piketower\flash\config\docker.cfg and change

DOCKER_ENABLED="yes"

to

DOCKER_ENABLED="no"

save, shutdown, and reboot.

I disabled the docker, and it boots and starts the array and stays stable, no crash.

 

I have the one drive that was unassigned when I tried starting the array with it unplugged, this drive shows as "missing" because the array is started.

 

Should I stop the array and re-build?  Then how do I get my Plex docker unistalled so I can re-install it?

 

Thank you!

I disabled the docker, and it boots and starts the array and stays stable, no crash.

 

I have the one drive that was unassigned when I tried starting the array with it unplugged, this drive shows as "missing" because the array is started.

 

Should I stop the array and re-build?  Then how do I get my Plex docker unistalled so I can re-install it?

 

Thank you!

At this point, you have 2 options. Rebuild, or set a new config and reassign all drives, check to trust parity, and do a full correcting parity check. If it were me, I'd do the new config.

 

If that's the option you want to take, make sure you know which drive goes in which slot by serial number before you set the new config. A screenshot of the main page is good.

 

After you get the array all green one way or the other, you need to delete the docker image file on disk1. Then you can let the docker manager create a new docker image file on the docker settings page and enable dockers again, and on the docker page add a new container, selecting the my-plex from the select a template drop down.

I have taken a screenshot of the main page, and a backup of my flash.

 

What are the steps to set a new config with trusted parity?

 

*edit*

 

Got it figured, it was under Tools.  I now have a correcting parity check running.

 

Tomorrow when the parity check completes, to delete the docker image file I just navigate manually to that drive and the docker directory and deleter it?

Got it figured, it was under Tools.  I now have a correcting parity check running.

 

Tomorrow when the parity check completes, to delete the docker image file I just navigate manually to that drive and the docker directory and deleter it?

Any parity errors so far? I'm hoping for single digits or zero.

 

To delete the file, yes, if you have disk1 exported as a disk share you can delete docker.img from the root of disk 1. If not, the easiest way would be to telnet in or at the local console type

rm /mnt/disk1/docker.img

It's not in a docker directory, it's in the root of the drive.

312 parity errors corrected, and they all came within the first minute of the parity check.

 

I have deleted the docker image file as well, next is to re-enable docker image and install Plex.

 

Thanks for the help!