Docker Compose plugin blank page with either chrome/firefox. Unraid 7.0 rc 1

 

1 hour ago, Jaffo73 said:

Docker Compose plugin blank page with either chrome/firefox. Unraid 7.0 rc 1

 

Looks to me like the first screenshot is of the APPs page not the compose page. Is there a problem with APPS as well?

the first one is chrome. i was on apps page and clicked on 'Compose' tab, not sure why the underline doesn't move (does now) but both chrome and firefox present the same blank page

 

Resolved. I had to make a new docker/compose-files directory and edit the plugin settings and point it there for compose projects. not shure why but its good now.

Sorry for what is I'm sure a very simple/stupid question. I ran a compose for TeslaMate without changing the volume paths to my appdata folder and now for the life of me I can't find the volumes so I can relocate them.  Am I missing something obvious about where these are stored I can't find them in the unraid file system.

 

- ./import:/opt/app/import

 - teslamate-db:/var/lib/postgresql/data

 - teslamate-grafana-data:/var/lib/grafana

  - mosquitto-conf:/mosquitto/config
  - mosquitto-data:/mosquitto/data

 

Thanks

 

13 hours ago, feta said:

Am I missing something obvious about where these are stored I can't find them in the unraid file system.

 - teslamate-db:/var/lib/postgresql/data

 - teslamate-grafana-data:/var/lib/grafana

  - mosquitto-conf:/mosquitto/config
  - mosquitto-data:/mosquitto/data

These are not exactly in the unRAID filesystem. The 4 above use volume syntax so docker volumes were created to back those mount points. They are named teslamate-db, teslamate-grafana-data etc. You should be able to see them using the command 

docker volume ls

You would also need to use docker volume commands to delete them or copy data out of them.

Hello,

for some months I am having a ghost container issue, I do not really know if the issue is unraid itself or this plugin

 

long story short before I was using immich with an all in one image, then I used this plugin to replace It with the docker compose setup

 

the issue is that this all in one image (imagegenius/immich) still reappears anytime I click upgrade all, when this happens my actual container (coming from docker compose) gets stopped and disappears, I have to click compose up and then gets recreated (you can see it getting recreated outside the media library folder, the one with the long string as docker network, idk why that happens to some containers and does not happen to others)

 

I do not know where in my system this cursed image is stored and the only thing in common between the old immich image and the (official) new one is the two having the same container name

 

thank you all

 

 

 

2 hours ago, mich2k said:

 

the issue is that this all in one image (imagegenius/immich) still reappears anytime I click upgrade all 

Unraids built in container update mechanism does not work correctly for containers not created by the native unraid interface (dockerman). When you have containers created by the compose plugin (or other non native managers) you should not use the update all button.

50 minutes ago, primeval_god said:

Unraids built in container update mechanism does not work correctly for containers not created by the native unraid interface (dockerman). When you have containers created by the compose plugin (or other non native managers) you should not use the update all button.

thank you for the hint

 

the interesting thing is that that ghost container was created through the GUI

Delete it, go to apps->previous apps, delete the immich template from there, then bring the stack up again. Issue is likely that the same name was reused so the UI thinks the contaner with the same name as the existing template go together when they don't.

Edited December 27, 20241 yr by Kilrah

9 hours ago, primeval_god said:

Unraids built in container update mechanism does not work correctly for containers not created by the native unraid interface (dockerman). When you have containers created by the compose plugin (or other non native managers) you should not use the update all button.

While the built in mechanism does not update non-dockerman containers, it also does not interfere with them. Clicking update all simply skips updating non-managed containers. I use it all the time as I have a mix of doxkerman and composeman stacks

 

The issue here looks like the old container is not removed from dockerman. I would bring the stack down, then update all, then properly remove the old container first (from apps/previous apps), before bringing the stack up. 

It seems that docker-compose manager is not updating the containers even if it should.... or that the web interface wrongly reports them as outdated.

 

PS. I am new to unraid forum,... I am still trying to understand how a forum is supposed to help as it seems a nightmare to navigate and find something in it,... i was expecting an issue tracker, to read existing open bugs...  The fact that on each search I get results that are like 4-5 years old first is... mind blowing. Who knows, maybe I am holding it the wrong way...

2 hours ago, ssbarnea said:

or that the web interface wrongly reports them as outdated.

Update status is not working for containers that are not managed via templates, ignore it.

22 hours ago, Kilrah said:

Update status is not working for containers that are not managed via templates, ignore it.

I submitted a PR a while ago but it hasn't been merged yet. It fixes this issue.

https://github.com/dcflachs/compose_plugin/pull/27

2 hours ago, mtongnz said:

I submitted a PR a while ago but it hasn't been merged yet. It fixes this issue.

The plugin author has previously mentioned he didn't want to take the risk of his plugin touching unraid's db so sadly it's unlikely it will be. 

Edited January 7, 20251 yr by Kilrah

I have a stack on which I clicked 'Update stack' a few weeks ago, and I accidentally closed the windowed that popped up in the middle of it doing those updates. After that I couldn't interact with the buttons for the stack at all, at least from the Docker Containers page, and there was a permanent Unraid loading icon in the center of the screen. I tried disabling Docker on Unraid and re-enabling, which worked for me to interact with those buttons one time, but after that the loading icon was back. Tried removing this compose manager tool, and reinstalling, but even still there's the loading icon when on that page. I figured out that I can access the buttons from the tool's own settings page, but surely there must be something I can clear/reset somewhere to fix this? 🙂

 

On Unraid 6.12.12, and Docker Compose Manager version 2024.09.13.

7 hours ago, emsixteen said:

I have a stack on which I clicked 'Update stack' a few weeks ago, and I accidentally closed the windowed that popped up in the middle of it doing those updates. After that I couldn't interact with the buttons for the stack at all, at least from the Docker Containers page, and there was a permanent Unraid loading icon in the center of the screen. I tried disabling Docker on Unraid and re-enabling, which worked for me to interact with those buttons one time, but after that the loading icon was back. Tried removing this compose manager tool, and reinstalling, but even still there's the loading icon when on that page. I figured out that I can access the buttons from the tool's own settings page, but surely there must be something I can clear/reset somewhere to fix this? 🙂

 

On Unraid 6.12.12, and Docker Compose Manager version 2024.09.13.

Could you provide some screenshots or something? I am not really sure what you are describing, particularly this part " I figured out that I can access the buttons from the tool's own settings page"

Here's a screen recording showing. Loading icon and can't interact on the Docker tab, then can access settings through the tool's own settings page.

6 hours ago, emsixteen said:

Here's a screen recording showing. Loading icon and can't interact on the Docker tab, then can access settings through the tool's own settings page.

Screen Recording 2025-01-13 at 10.04.16.mov

Interesting, that seems to suggest that its the Dockerman tab that is hanging rather than the compose tab specifically. Does bringing the stack down change the behavior at all?

Side note I was not aware that the setting link in CA was wrong. It should lead to the compose plugin setting page rather than the compose page.

Edited January 13, 20251 yr by primeval_god

On 1/13/2025 at 4:16 PM, primeval_god said:

Interesting, that seems to suggest that its the Dockerman tab that is hanging rather than the compose tab specifically. Does bringing the stack down change the behavior at all?

Bringing the stack down doesn't fix the issue – The loading icon's still there, and I can't interact with the buttons along the bottom of the Docker Containers section, or the buttons in the Commands column added in the Compose section.

 

I just found what the issue what the issue was though, right after writing that. I'd removed one of the applications from the stack, recyclarr, and it was still in the Docker Containers list saying Rebuilding. Can be seen briefly in my video. Removing that from there fixed the issue. 🙂

 

---

 

I have another little question actually. When I use update stack everything seems to update, but in the list they all still say "Apply update" - I'm guessing this is normal?

Edited January 15, 20251 yr by emsixteen

3 hours ago, emsixteen said:

I have another little question actually. When I use update stack everything seems to update, but in the list they all still say "Apply update" - I'm guessing this is normal?

This is normal, the reason has been discussed previously in this thread.

 

Hi, have noticed a couple of things since upgrading to unraid 7.

• The 'Docker' tab in the UI doesn't show correctly that containers are part of a compose stack.  Some do and some dont but all my containers belong to 1 of 6 stacks.  I've re-saved the compose file via the UI to regenerate the docker-compose.override.yml files in the projects directory but the below is still happening, obviously no actual functionality issues and all cosmetic.  

• I've started to receive this message when starting a stack, is this the compose pluging or docker where I need to research?

matter Migrated sysctl "net.ipv6.conf.eth0.accept_ra_rt_info_max_plen" to DriverOpts{"com.docker.network.endpoint.sysctls":"net.ipv6.conf.IFNAME.accept_ra_rt_info_max_plen=64"}

 

Any pointers appreicated.

 

Thanks K

2 hours ago, KillerK said:

• The 'Docker' tab in the UI doesn't show correctly that containers are part of a compose stack.  Some do and some dont but all my containers belong to 1 of 6 stacks.  I've re-saved the compose file via the UI to regenerate the docker-compose.override.yml files in the projects directory but the below is still happening, obviously no actual functionality issues and all cosmetic. 

The display elements you are referring to are not part of the compose plugin, I dont really know how they work. The "Compose" vs "3rd Party" tag is controlled by the `net.unraid.docker.managed=composeman` label on each container, which should be applied as part of the compose.override.yml file. You can use docker inspect from the cli to check what labels are applied to each container. Can you verify that the compse.override.yml (for each stack) file has the above label set for each container?

 

2 hours ago, KillerK said:

• I've started to receive this message when starting a stack, is this the compose pluging or docker where I need to research?

matter Migrated sysctl "net.ipv6.conf.eth0.accept_ra_rt_info_max_plen" to DriverOpts{"com.docker.network.endpoint.sysctls":"net.ipv6.conf.IFNAME.accept_ra_rt_info_max_plen=64"}

This is likely something from Docker or Docker Compose (the program not the plugin). 

1 hour ago, primeval_god said:

The display elements you are referring to are not part of the compose plugin, I dont really know how they work. The "Compose" vs "3rd Party" tag is controlled by the `net.unraid.docker.managed=composeman` label on each container, which should be applied as part of the compose.override.yml file. You can use docker inspect from the cli to check what labels are applied to each container. Can you verify that the compse.override.yml (for each stack) file has the above label set for each container?

 

So one of my stacks is called 'other_apps' because I'm a creative type 🤔  It has 5 containers described, 3 of which correctly report as part of the stack on the 'Docker' UI page but 2 don't.  Interestingly or not the 2 that don't are the last 2 additions (and the last 2 in the below insert) but were added a while back and not together at the same time.

 

# cat /boot/config/plugins/compose.manager/projects/other_apps/docker-compose.override.yml
services:
  myadmin:
    labels:
      net.unraid.docker.managed: 'composeman'
      net.unraid.docker.icon: ''
      net.unraid.docker.webui: ''
      net.unraid.docker.shell: ''
  esphome:
    labels:
      net.unraid.docker.managed: 'composeman'
      net.unraid.docker.icon: ''
      net.unraid.docker.webui: ''
      net.unraid.docker.shell: ''
  bab_wordpress:
    labels:
      net.unraid.docker.managed: 'composeman'
      net.unraid.docker.icon: ''
      net.unraid.docker.webui: ''
      net.unraid.docker.shell: ''
  pgadmin:
    labels:
      net.unraid.docker.managed: 'composeman'
      net.unraid.docker.icon: ''
      net.unraid.docker.webui: ''
      net.unraid.docker.shell: ''
  netbootxyz:
    labels:
      net.unraid.docker.managed: 'composeman'
      net.unraid.docker.icon: ''
      net.unraid.docker.webui: ''
      net.unraid.docker.shell: ''

 

There is a clear difference you can see via docker inspect also...

 

# docker inspect mysql | grep net.unraid
                "net.unraid.docker.icon": "",
                "net.unraid.docker.managed": "composeman",
                "net.unraid.docker.shell": "",
                "net.unraid.docker.webui": ""
# docker inspect pgadmin | grep net.unraid
# docker inspect netbootxyz | grep net.unraid

 

Now I am guilty of bypassing the UI and using SSH to administer my containers including updating the compose files and starting/downing stacks.  However even after manually renewing the config via the UI this isn't being corrected unless there is some linkage to a reboot I've not yet tried?

Edited January 21, 20251 yr by KillerK

1 minute ago, KillerK said:

No I am guilty of bypassing the UI and using SSH to administer my containers including updating compose files and starting/downing stacks.  However even after manually renewing the config via the UI this isn't being corrected unless there is some linkage to a reboot I've not yet tried?

No reboot should be necessary but you do have to run Compose Up after making any changes for them to take effect. I dont think it is required but have you tried running Compose Down and then Compose Up to recreate the stack?