Docker list is really slow on 6.8.1

[mikamap]

  Hi,

 

  I'm currently on 6.8.1 with an unraid OS Plus licence.  Right now, I have a problem when i click on the docker tab in the web interface.  I tried with chrome, IE, firefox and on my cellphone and the docker list is still really slow to show.  It take more than a minute to display the list of docker.

 

  I presently have only 8 docker image and 1 VM running.

  Even when I stop all docker, the list is really slow

 

  I first noticed it when I upgraded to 6.8.1.

  This is what I did for testing :

 

  1- Revert back to previous version. 

  I think it was 6.7.2.  Problem was resolved.  The list was showing fast.

  I updated back to 6.8.1 because I though the previous update was not installed correctly and the problem was quickly back.

  After that, I went back to 6.7.2, but the docker list was slow again.

  So I upgraded back to 6.8.1.

 

  2- Change hardware.

  I previously had a intel 2 core and most of the time, the CPU was at 100%.  So i tought it was the hardware.

  So I just upgraded to AMD 2700X with a brand now motherboard and 16Go or RAM.

  The docker list is still slow (As much as always) even with only 4% of CPU usage.

 

  I know I have SMART errors on 1 off my disk at the moment, can It be the culprit ?

 

  I check the system log and did not find anything weird.  I uploaded my diagnostics that I just made with the new hardware.

  The only thing I did sine then is delete 2 plugins : networkstats.plgand openvpn_server_x64.plg

 

  When I put Network logs in firefox, I can see that the DockerContainers.php is the one taking a long time.  But, I cannot see error in firefox console log.

 

 

  Is there a way to activate a debug log to see what is happening when the docker list loads ?

 Or to find what does DockerContainers.php actually does to try to find the problem ?

 

 

  Thank you for your help.

 

Edited February 1, 2020 by mikamap
Problem was corrected

[bonienl]

Try the following

 

1. Stop the docker service

2. Delete the docker image

3. Create a new docker image, highly recommended on the cache device

4. Re-install your containers (use CA to do this)

[Squid]

Curious when you on 6.8.x what is the output of this command?

 

ls /var/lib/docker/unraid/images

Is it showing an icon for every container you have installed?

[mikamap]

2 hours ago, bonienl said:

Try the following

 

1. Stop the docker service

2. Delete the docker image

3. Create a new docker image, highly recommended on the cache device

4. Re-install your containers (use CA to do this)

Just to be sure, I won't loose any settings for the dockers ?  I don't want to reupload everything to crashplan after that.

 

Thanks

[Squid]

1 minute ago, mikamap said:

Just to be sure, I won't loose any settings for the dockers ?  I don't want to reupload everything to crashplan after that.

 

Thanks

https://forums.unraid.net/topic/57181-docker-faq/#comment-564363

 

[mikamap]

2 hours ago, Squid said:

Curious when you on 6.8.x what is the output of this command?

 

ls /var/lib/docker/unraid/images

Is it showing an icon for every container you have installed?

No it's not.  It's missing 2 images.

 

I deleted the 2 docker missing the images and now the docker list is super fast again.

 

Thanks.  And it was docker I did not use for a long time.  I'll reinstall them if I need them in the future.

[Squid]

Just now, mikamap said:

No it's not.  It's missing 2 images.

 

I deleted the 2 docker missing the images and now the docker list is super fast again.

 

Thanks.  And it was docker I did not use for a long time.  I'll reinstall them if I need them in the future.

Which 2?

 

An underlying change in 6.8+ was that icons needed to be downloaded again the first time you hit either the dashboard or the docker tab.

[mikamap]

1 minute ago, Squid said:

https://forums.unraid.net/topic/57181-docker-faq/#comment-564363

 

Thank you.  I don't need to do this as deleting 2 dockers without images did the trick for me. 

But As you recommended by bonienl post, I'll put the docker image on the cache (I did not have a cache when I first installed unraid and used docker)

[mikamap]

8 minutes ago, Squid said:

Which 2?

 

An underlying change in 6.8+ was that icons needed to be downloaded again the first time you hit either the dashboard or the docker tab.

One of them was a custom docker I made.  it was hosting a Spring boot java application.  It was something I made to practice for a job interview.

 

And the other one was a docker I installed from a custom repo.  It was a spring boot admin instance.  I think the repo from which I installed it was removed or made private as I could not find it anymore.

 

Bottom line, I'll stick with CA application from now on

Edited February 1, 2020 by mikamap

[Squid]

Can you do me a favour and post or PM me the applicable xml files from /config/plugins/dockerMan/templates-user

[mikamap]

4 minutes ago, Squid said:

Can you do me a favour and post or PM me the applicable xml files from /config/plugins/dockerMan/templates-user

  Oh, I think I know what the problem was.

 

  When I made the docker, I hosted the image on my gitlab instance (hosted of unraid).

  But when I change router, my ip address changed so the IP address of unraid also changed.  So the image was not available from this point on.  It was a few months ago, but when I updated to 6.8.1, it broke my docker list.

 

 I uploaded the xml for the application.

You'll see 192.168.2.116, but it is not my unraid ip anymore.

 

 Thank you

 

 

my-SinopeApiMQTT.xml

[Squid]

A bit irrelevant now, but I'll issue a fix for the next unRaid version that will lower the connect timeout to 15 seconds, rather than the combined 60 seconds for download & connect.

 

Effectively though, the issue you're seeing only happens in your self-inflicted circumstance (no offense meant), as any URL would basically return a 404 immediately.

[mikamap]

2 minutes ago, Squid said:

A bit irrelevant now, but I'll issue a fix for the next unRaid version that will lower the connect timeout to 15 seconds, rather than the combined 60 seconds for download & connect.

 

Effectively though, the issue you're seeing only happens in your self-inflicted circumstance (no offense meant), as any URL would basically return a 404 immediately.

No offense taken.  I know the problem is all my fault.

 

I'm really grateful for the help.

 

[Squid]

Full disclosure though, I pumped through changes to how the icons were handled in 6.8 so that Joe Blow user could change the Icon URL at any given time and unRaid would pick up the changes properly (it never really did previously without jumping through a lot of hoops), but a downside was the system had to re-download everything the first time it hit dashboard / docker.  I knew there would be an edge case or 2 where this would cause an issue, but you win the prize for being the first...

 

 

[richardsim7]

Am I the second?

 

Found this thread after noticing my Docker page was causing the page/my browser to be slow. Found an extra image after changing hydra2 to nzbhydra2 as per linuxserver's instructions. Deleted it and things appear to be back to normal (I hope)