Docker self deleted after adding path failed

[jnevada]

Hi,

 

I installed the official Prestashop Docker and after configuring for some hours I wanted to acces files in the docker manually so I tried adding a path from / to a new folder on Unraid (/mnt/users/appdata/prestashop, I checked before and it didn't exist).

 

This failed and the docker disappeared. After reading some posts I tried reinstalling from previous apps but all data is gone and it behaves like a new installation.

 

Where is the original docker gone? Did it really self-delete after trying to add a path or is it sitting somewhere as a ghost? Diagnostics:

http://www.filedropper.com/greed-diagnostics-20210316-2214

 

Regards

 

[trurl]

I don't use that docker but you said its appdata didn't exist so I would expect it to be a new install. 

 

On the Docker page change from Basic to Advanced with the slider at upper right and see if you have any orphans. Nothing you can do but delete them. 

 

Why did you want access inside the container? You can get a command line inside any container by clicking its icon and selecting Console. 

[Squid]

2 hours ago, jnevada said:

After reading some posts I tried reinstalling from previous apps

Previous apps works by installing from the last template.  IE: it has the same error in it.

 

On any failed install, the docker does "disappear", because it failed to install.  Just fix the error and it's back.

 

A path not existing won't cause a failed install, because the system will automatically create the path.  The docker run command that was shown after you did this would have given the exact reason why the container failed to install.  You can get it again by simply doing previous apps and hitting the download icon from the tile (NOT checking it off and then hitting install)

[jnevada]

4 hours ago, trurl said:

I don't use that docker but you said its appdata didn't exist so I would expect it to be a new install. 

 

On the Docker page change from Basic to Advanced with the slider at upper right and see if you have any orphans. Nothing you can do but delete them. 

 

Why did you want access inside the container? You can get a command line inside any container by clicking its icon and selecting Console. 

Because the console in the docker behaves completely different than the unraid one. There are no orphans. I don't understand where the docker is gone.

 

3 hours ago, Squid said:

Previous apps works by installing from the last template.  IE: it has the same error in it.

 

On any failed install, the docker does "disappear", because it failed to install.  Just fix the error and it's back.

 

A path not existing won't cause a failed install, because the system will automatically create the path.  The docker run command that was shown after you did this would have given the exact reason why the container failed to install.  You can get it again by simply doing previous apps and hitting the download icon from the tile (NOT checking it off and then hitting install)

Not sure if I use the terms correctly but the installation did not fail initially. It was running and I setup everything as I needed. Stopping, starting restarting all worked.

 

Then I literally only added a path with no other changes. And I know it creates the path and it did but the previously running docker failed and deleted due to this single change.

 

Normally when I changed settings and they were "wrong" they simply were unstartable and I could change/fix the settings. On this one the docker "disappeared", I mean from the list of dockers and was only to be found in previous apps and there are no orphans.

 

The reinstall from previous apps works with the settings that caused the deletion, but all the data is gone and it is in a state as if I created a brand new docker. So my question is where is the original docker gone? Did it delete? Is it sitting somewhere as a ghost?

 

I did look at the logs (which I uploaded) but I don't understand what happened and why, hence I created this thread:

time="2021-03-16T13:52:39.539972602-07:00" level=error msg="Handler for POST /v1.40/containers/create returned error: invalid volume specification: '/mnt/user/appdata/prestashop:/:rw': invalid mount config for type \"bind\": invalid specification: destination can't be '/'"

 

I'm trying to find out how to prevent this from happening again. As it is not ideal to lose all development without knowing why and if it will happen again.

Edited March 18, 2021 by jnevada

[trurl]

3 hours ago, jnevada said:

the logs (which I uploaded)

I haven't looked at those since they are at an external site. Attach your diagnostics directly to your NEXT post.

[Squid]

7 hours ago, jnevada said:

invalid specification: destination can't be '/'"

Edit that path mapping.  You're mapping it to /, which isn't allowed.  you need to give it a folder name within the container.  (possibly /config)

[tjb_altf4]

if a run command fails due after editing, you can actually go back with the browser controls and you can fix your mistake and resubmit

[jnevada]

7 hours ago, Squid said:

Edit that path mapping.  You're mapping it to /, which isn't allowed.  you need to give it a folder name within the container.  (possibly /config)

I got that after it failed, the issue is there is no docker left to edit. If adding this path causes the deletion of the docker I would say that is a very simple bug to fix with an error message "path to / not allowed" without ever applying it but surely the docker shouldn't be deleted even if a wrong setting is applied.

 

11 hours ago, trurl said:

I haven't looked at those since they are at an external site. Attach your diagnostics directly to your NEXT post.

Sorry I didn't know that I can upload here, do I have to make a new thread or is this enough? Basically after failed mounting the docker got marked for removal is what I gather.

 

6 hours ago, tjb_altf4 said:

if a run command fails due after editing, you can actually go back with the browser controls and you can fix your mistake and resubmit

The docker got removed after the failed config, I could see "this container does not exist". I don't think it was simply disappearing visually. The docker got marked for removal and was gone after failing to mount the "wrong" path.

greed-diagnostics-20210316-2214.zip

[trurl]

11 hours ago, jnevada said:

do I have to make a new thread

No in fact I usually say "NEXT post in this thread" to avoid confusion but didn't that time. "NEXT post" to make sure the thread shows up as having new content, sometimes people will edit a previous post without making a new post and nobody will know there is anything else to see.

 

Unfortunately diagnostics often don't tell us much about specific dockers, but your overall docker configuration looks good.

 

When you edit a docker and submit that page, it has to remove the docker so it can create it again with the new mappings, etc. In your case, since the attempt to create it again failed due to the illegal mapping, the docker no longer existed.

 

[Squid]

14 hours ago, jnevada said:

I got that after it failed, the issue is there is no docker left to edit. If adding this path causes the deletion of the docker I would say that is a very simple bug to fix with an error message "path to / not allowed" without ever applying it but surely the docker shouldn't be deleted even if a wrong setting is applied.

It's not actually deleted.  It's appearing as an Orphan if you click Advanced View.  But, you fix it via Previous Apps, hit the download icon for that app, and edit the path accordingly (even if you have to hit show more settings)

 

And, the orphan / deleting however you want to word it is how the entire docker system works across the board on all OS.  Many reasons for the docker "run" to fail, and it will always result in an orphan because the system can't actually run it with the parameters you gave it.