[BUG] Dashy crashlooping after last upgrade to 2.1.2

[Matsched]

[BUG] Dashy crashlooping after last upgrade to 2.1.2

[Lissy93]

Okay, updated version live on DockerHub and GHCR. Try re-pulling :latest or :2.1.2 tag.
It's working for me and @CrazyWolf13 - but OP give it a shot and let us know how you get on.

And sorry for the trouble this has caused, I'll do smaller and more frequent updates going forward, to hopefully mitigate things breaking.

[CrazyWolf13]

It stops working, how I can check what is a reason and share here?

Lissy has already identified the issue and is currently testing a fix.

Huge claps to her 👏

[Lissy93]

Sorry, sorry, yes fix coming in #1496
Just testing it at the moment.....

The issue was a bug causing loads of rebuilds, causing it to consume more ram than intended. I've written a new re-build script which should only watch for relevant changes... just testing it on several devices.

[Lissy93]

Merging #1496 now....
Will be available on DockerHub shortly
I'm hoping this will fix the issue, so 🤞
Let me know if that works out for you

[CTHRU]

This might explain my issue too on a rpi 3b docker installation. The rpi runs the latest 64 bit pi os lite version, fully upgraded. I noticed that v2.1.2 slowly uses all assigned memory resources (448MB) after a few minutes. It also uses all assigned cpu resources (50%) while doing so.

I have a slightly adapted docker file where i bind mount the whole app/public folder. And yes, i did copy the folder's content from the v2.1.2 source code on GitHub on the host folder.

compose.yaml file

services:
  dashy:
    container_name: dashy
    hostname: dashy
    memswap_limit: 448M
    deploy:
      resources:
        limits:
          cpus: "0.50"
          memory: 384M
    restart: unless-stopped
    ports:
      - 80:80
    image: lissy93/dashy:latest
    volumes:
      - ${HOST_APP_DIR}/public:/app/public
    networks:
      - dashy
networks:
  dashy:
    name: dashy

• I saw the docker hub update of this afternoon and pulled the image again, allas same result.
• the default conf.yml website is loaded long before the memory is consumed.
• I saw the fancy new loading screen (nicely done).
• I tried to clear the site's browser data but the behavior is consistent.

[Matsched]

I have tested a bit:
If i docker compose up dashy and immediately open the dashboard in browser i get

dashy  | Error: ENOENT: no such file or directory, stat '/app/dist/index.html'

And Dashy exit with errorcode.

After disabling resource-limits in docker-compose.yml and waiting for dashy to fully start up, dashy is working. I can repeat F5 without any issue.

If i open Dashy on unlimited resources before it fully starts up, i get again "ENOENT: no such file..." like above and dashy hang up.

Now i see in: https://dashy.to/docs/deployment/ Dashy needs 1GB. It runs all the time on 256MB flawlessly (lol)

[Lissy93]

Now i see in: https://dashy.to/docs/deployment/ Dashy needs 1GB. It runs all the time on 256MB flawlessly (lol)

Nothing changed there, those were just the recommended specs for the Node base image, but it very very very rarley gets anywhere near 1GB

[liss-bot]

If you're enjoying Dashy, consider dropping us a ⭐
_{🤖 I'm a bot, and this message was automated}

[Lissy93]

Hmm, that's not good.
I did push an update a few hours ago, was this before or after that update?
Could you try re-pulling that latest image / tag?

[ulrik-vj]

Also having this issue. I can run the container as root (2.1.2/latest) no problems. But running as user: XXXX:XXXX, gives constant crashing.
Just pulled and Tested on 2.1.1, there i can run as user no crashing. Here is something from logs running latest image:

[CrazyWolf13]

Hmm, that's not good. I did push an update a few hours ago, was this before or after that update? Could you try re-pulling that latest image / tag?

Actually getting the same error :/

[p6002]

It stops working, how I can check what is a reason and share here?

[p6002]

Thanks it works now!
sudo docker pull lissy93/dashy:latest