I am having an issue where my system seems to run just fine, and then, poof, GUI becomes inaccessible.

 

I've seen what seem to be two flavors of this.

 

The first seems to involve Docker. I'll try to update or stop a container, and it will hang. From then, GUI is inaccessible. However, the GUI of any running dockers respond perfectly normally. Sometimes I can use the CLI to kill or stop the docker processes and get myself through a clean reboot. Sometimes I'm having to do a 'shutdown -n -r now' to get the system to actually reboot. I know this isn't a good thing, but apart from an actual forced power down I don't see any other way around it. As near as I can remember, I have seen this happen ONLY when I had just tried to do something to a Docker container. Which means I can (generally) avoid it, as long as I'm not stopping or updating a container via CLI.

 

When I have this first issue, when I try to access my unRAID box I get an http connection refused.

 

The second flavor of losing the GUI is just...losing the GUI. It happened yesterday around 2040PM or so, when the system hadn't been up for even an hour yet. This is the one I've got the syslog attached for.

 

When I have this second issue, I just get http timeouts when I try to access the box.

 

From a performance perspective, the only thing I've really noticed on the system is that if I (going by memory here since I have no GUI at the moment) go under applications, and try to use one of the default searches, it throws an error that lists about 5 lines of code. Which is obviously not all that helpful without my having the code to share.

 

I've actually got time to work on this issue today, as opposed to my usual schedule, so I am actually try to chase answers instead of just flipping the switch here.

 

Any help is greatly appreciated.

unraid-diagnostics-20160508-0726.zip

Sometimes I'm having to do a 'shutdown -n -r now' to get the system to actually reboot. I know this isn't a good thing, but apart from an actual forced power down I don't see any other way around it.

Am unable to assist right now with the actual issue, but you should install the powerdown plugin (available in CA or at http://lime-technology.com/forum/index.php?topic=31735.0) and instead of shutdown commands, issue

powerdown -r

  far safer to do it this way

I actually do have that installed - and in a situation like this (where things have gone funny) it doesn't seem to have help.

 

For example, to see if I had it installed (I thought I had, wanted to make sure), I just ran powerdown, expecting a list of command line switches.

 

And I get this:

 

root@UnRAID:~# powerdown
Powerdown initiated
mkdir: cannot create directory ‘/boot/logs/’: Input/output error
Powerdown V2.20
mkdir: cannot create directory ‘/boot/logs/’: Input/output error
Capture diagnostics to /boot/logs
Starting diagnostics collection... mkdir: cannot create directory ‘/boot/logs’: Input/output error
done.
ZIP file '/boot/logs/unraid-diagnostics-20160508-0746.zip' created.
mkdir: cannot create directory ‘/boot/logs/’: Input/output error
mkdir: cannot create directory ‘/boot/logs/’: Input/output error

Broadcast message from root@UnRAID (pts/0) (Sun May  8 07:46:53 2016):

The system is going down for system halt NOW!
^C

 

I didn't have the USB key in, which explains the error messages you see. After 15 minutes I have finally lost SMB, but it's still pinging away.

 

 

 

Why do you not have the usb key in? That could be the cause of all your issues.

I did not have the USB key in so that I could move it to this PC so that I could upload the syslog. When I ran the powerdown command I forgot that I had had the USB key in this machine, rather than the unRAID box.

Why not simply copy the log file from the flash share via SMB?

Because I didn't think of it.

 

Also, I'm now two hours into the powerdown script. The drives are presumably unmounted, since none of the dockers or SMB work anymore. I also can't telnet at this point.

 

However, still pinging away.

Because I didn't think of it.

 

Also, I'm now two hours into the powerdown script. The drives are presumably unmounted, since none of the dockers or SMB work anymore. I also can't telnet at this point.

 

However, still pinging away.

After two hours, I would have given up on waiting for it to shutdown.

I did, just before you posted.

 

So, I bounced it. And for the hell of it I put myself on the newest beta.

 

Is there anything I can be running to determine what exactly is killing the GUI?

 

...also, I've now got a screenshot of the weirdness with Apps.

Update CA

Alrighty, updated CA and that took care of that. There was an update available for Powerdown so I got that updated as well.

Alrighty, updated CA and that took care of that. There was an update available for Powerdown so I got that updated as well.

Now you can go to settings / CA Settings / Autoupdate settings and pick and choose which plugins to automatically keep up to date.

That's done. I'm now updating my dockers one at a time. I believe it'll be binhex's deluge-vpn that will make things go berserk, if anything does, so I'm saving that one for last...

Well, the good news is that I managed to update everything apart from binhex-delugevpn.

 

The somewhat good news is that an attempt to update binhex-delugevpn doesn't seem to have killed my web GUI.

 

The bad news is, when I try to update binhex-delugevpn (which is there and functional), I get: Configuration not found. Was this container created using this plugin?

Well, the good news is that I managed to update everything apart from binhex-delugevpn.

 

The somewhat good news is that an attempt to update binhex-delugevpn doesn't seem to have killed my web GUI.

 

The bad news is, when I try to update binhex-delugevpn (which is there and functional), I get: Configuration not found. Was this container created using this plugin?

Only seen that error once.  And I think my solution was to kill docker.img and start it over again.

So, assuming I do that, what exactly do I lose?

 

Fact of the matter is that all the possible directories are various user shares...so what do I lose when I wipe and redo the docker.img?

So long as you weren't storing anything within the docker.img (and only if you configured everything wrong would you be) everything (all settings, etc) are all stored within appdata.  After toasting the image, recreating it, and going to CA's previous apps section and adding the containers back in, after the download finishes you won't even be able to tell the difference.  It'll be the exact same as before.

That's what I thought. Thanks for all your help. Maybe I'll be super lucky and the beta will resolve my randomly quitting GUI. I can hope, anyway.

Try this plugin:

https://lime-technology.com/forum/index.php?topic=47386.msg469462#msg469462

it fixes a problem in smartmontools which caused at least one person to lose access to their unraid web gui.