Out of space problem

[Woz32]

Hi,

 

I'm a bit of a newbie so apologies!

 

I have a UnRaid server with 13.5TB across 5 drives and a 500GB SSD cache. I stupidly put too much on my array and have about 50MB (megs!) free and 200gig free on my cache. I now can't start the array to free some space or access the drives. 

 

Does anyone have any suggestions? I know its a bit of a stupid error but I was busy and not paying attention.

 

Log attached

 

Thanks,

Waz

server-diagnostics-20181111-1936.zip

[itimpi]

Lack of space should not stop the array from starting!

 

i could not see any evidence in the log of you attempting to start the array (Autostart is disabled).

 

Reading the log is difficult as it is full of messages from your Microsoft Optical Mouse about disconnecting and then reconnecting every 50 seconds or so.   You might want to try this in another USB port (I would suggest a USB2 port if you have one) to try and reduce the ‘noise’ in the log.    Also I would expect the continual USB disconnects/reconnects are likely to be adversely affecting performance.

[Woz32]

Thanks itimpi. Log without the USB issues attached.

 

Many thanks for your help

syslog.txt

[itimpi]

According to that log the array HAS started correctly.

 

it also shows that the loop2 image (which is the docker image file) is corrupt so docker containers will not be running.   You can correct this by stopping the docker service; deleting the docker image file;  restarting the docker service (which will create an empty docker image file); and then reinstalling any docker containers via the /Previously Installed apps option within Community Applications

[Woz32]

Strange, it definitely says Array starting...starting services. The docker tab has Array must be started to view Docker containers.

[Squid]

Diagnostics are always better than syslogs as there *may* be an underlying issue that cannot be determined purely from the syslog

 

Beyond that,

 

unRAID System Management Utility version 6.5.0

I can't think of a valid reason to not update the OS to at least 6.5.3 much less 6.6.5

[itimpi]

Sounds like the problem is starting the docker service then.     You could edit the docker.cfg file in config folder on the flash drive and change the line that says the docker service is enabled from “yes’  to “no”.   Hopefully that will then allow the startup to complete successfully so you can follow the steps for correcting the docker image file problem.

[Woz32]

Thanks for the help. I deleted the docker image file which did allow me to restart the docker services. However, I am still getting the overall services unable to start. Full diagnostics attached.

 

 

server-diagnostics-20181112-1130.zip

[itimpi]

I can see that the log file is full of error messages but no idea what is causing it?

 

Have you tried starting Unraid in Safe Mode in case it is a plugin that is causing your problem (Safe Mode suppresses any plugins being started).

[trurl]

Nov 12 11:17:57 Server emhttpd: error: put_config_idx, 609: Invalid argument (22): fopen: /boot/config/shares/?.cfg

You have a user share named _

That is, the underscore character is the name of that user share.

 

It's not obvious to me that there should be some reason that is not allowed, but it is unusual, and might correspond to the error in the log.

 

The user shares are just the aggregated top folders of cache and the array disks. Even if you don't intentionally create a user share in the webUI, if you have a top level folder on cache or array it is automatically a user share with the same name as the folder.

 

Do you know how you might have gotten a top level folder on cache or array named _

(the underscore character)?

[Woz32]

4 hours ago, trurl said:

Nov 12 11:17:57 Server emhttpd: error: put_config_idx, 609: Invalid argument (22): fopen: /boot/config/shares/?.cfg

You have a user share named _

That is, the underscore character is the name of that user share.

 

It's not obvious to me that there should be some reason that is not allowed, but it is unusual, and might correspond to the error in the log.

 

The user shares are just the aggregated top folders of cache and the array disks. Even if you don't intentionally create a user share in the webUI, if you have a top level folder on cache or array it is automatically a user share with the same name as the folder.

 

Do you know how you might have gotten a top level folder on cache or array named _

(the underscore character)?

Not sure, its actually called "?". It dosent have any data. Shall I just delete it?

[John_M]

A folder called ? would very likely cause problems because the question mark and asterisk are used as wildcards.

[trurl]

9 hours ago, Woz32 said:

Not sure, its actually called "?". It dosent have any data. Shall I just delete it?

yes

[Woz32]

15 hours ago, trurl said:

yes

 

Thanks, that did the trick!

 

Thanks for all the other help as well.