After booting the Pi, if the asset set is a single image, it doesn't display, the screen remains dark with only the small purple screenly logo in the center

Anthias
21

Hi @nicomiguelino !, Have these problems been resolved in the new March release by any chance?

22

@caimano, not yet, as the issue stopped persisting during the point when I updated some dependencies. Do you still experience the issue?

I’ll be reopening the following pull request if it still occurs: Fixes: Anthias is not starting asset (gets stuck at the logo screen) by nicomiguelino · Pull Request #1896 · Screenly/Anthias · GitHub

23

Hi, what dependencies have you update to solve this issue ?

24

I’ve upgraded the to Celery 5 (from 4). Do you still experience the issue?

25

Even the last version of Anthias does not work.
When I turn off and on the raspberry if I have only one image in the playlist, it is not loaded, I have to remove it and put it back manually. If instead there are more images in the playlist then they start automatically when turned on. I don’t know what Celery 5 is. Could you explain to me what it is and how I can also make this modification to make it work? Thanks a lot

26

Thanks, @caimano, I really thought it’s just a fluke, as it happens less than 50% of the time on my end when I was testing it. I’ll update the following PR, test it, have it for review, then merge it — Fixes: Anthias is not starting asset (gets stuck at the logo screen) by nicomiguelino · Pull Request #1896 · Screenly/Anthias · GitHub. I’ll let you know when it’s ready.

EDT: @caimano, Celery is a task queue implementation for Python. It’s used by Python programs to to tasks asynchronously. As the issue still persists, Celery has nothing to do with the fix/issue.

27

@caimano, we just shipped changes that might fix the issue. Make sure to backup your data before updating Anthias (if your using the Raspberry Pi OS version).

Let me know if you’re still having the issue.

28

We have been struggling with this over the past week or so with new builds on Pi 3 B+ and Pi 4 B units. To install we use Raspbian Lite, run updates and raspi-config, and run the Anthias install shell script.

Once configured and loaded with a single asset I was able to follow the process and see that it would get to the loading of the STANDBY_SCREEN, trigger the main loop, and trigger the asset (the asset is a web page if that is relevant), but that asset would not be displayed. If there was a second asset, it would be displayed after the delay for the first asset expired which would get things working. If we changed something about the single asset to cause the asset list to be generated again this would also resolve the issue.

To resolve this at the viewer.py level, I noted that there is a sleep delay built into the end of the asset_loop but there was no delay after the initial loading of the STANDBY_SCREEN in main. Once I added a sleep directly before the call from main to start_loop, everything started working as expected for single asset lists after reboots or restarting the viewer container.

Original viewer.py code:

    view_image(STANDBY_SCREEN)

    load_screen_displayed = True

    if mq_data is not None:
        show_hotspot_page(mq_data)
        mq_data = None

    start_loop()

Update to resolve single asset issue:

    view_image(STANDBY_SCREEN)

    load_screen_displayed = True

    if mq_data is not None:
        show_hotspot_page(mq_data)
        mq_data = None

    sleep(0.5)

    start_loop()
29

@QuantumHaze, nice catch! Thank you for sharing your solution. I’ll create a pull request later on and will keep you posted.

P.S. I really thought that the problem is solely in the C++ code.

30

@QuantumHaze, the fix was merged and shipped a few days ago. Please let me know whether you’re still having the issue or not.

1 Like
31

Tested this morning and everything is working as expected. Thanks for the quick turnaround!

1 Like
32

Awesome, thanks!
@QuantumHaze, please mark the following post as the solution, if possible:

33

Great!! So I can try to do a test now too. Thank you very much! But if I wanted to update my raspberry without having to reinstall and reconfigure everything can I do it via command line without losing the previous configurations? What is the command I should run? Then to be able to verify that the version of Anthias has actually changed where can I see it? Regards.

34

@caimano, I would recommend you to run the following:

bash <(curl -sL https://install-anthias.srly.io)

The install script also acts as the update script for existing installations.
If the version of your Anthias is quite old, I also recommend you to download the backup data by going to Settings and clicking Get Backup. It should download a tar.gz file.

35

@nicomiguelino, i tried to configure a static ip address to know if after the upgrade, the configuration was the same, but in this version i don’t find the file dhcpcd.conf inside /etc/ folder. There are some difference ? Can i set it by sudo raspi-config?

36

@caimano, what OS version does your Raspberry Pi Run? (Is it Bullseye or Bookworm?)

After quickly checking raspi-config, it’s not possible to do so.
If you’ve selected Yes for network management during the update process, then dhcpcd will be disabled in favor of NetworkManager.

Setting a static IP address (thru NetworkManager) can be done via nmcli or nmtui.

37

Thanks @nicomiguelino.
Finally it seems that the problem of the single image that did not start has been solved. I did the update with the bash command <(curl -sL https://www.screenly.io/install-ose.sh) and then I configured a fixed IP with nmtui as you advised me.

However, I discovered that after I restarted the raspberry when the update was finished, leaving it with a fixed image, after many hours, an error message appears at the top of the display like this: " mmc1 controller never released inhibit bits " followed by two other strings.

Is it an error generated by Anthias or by the Raspberry s/o? (I used Raspberry Pi OS (64-bit): 2024-03-15-raspios-bookworm-armhf-lite.img.xz).

I found something about it on the Raspberry forum (https://forums.raspberrypi.com/viewtopic.php?t=330384&sid=461468cf9e2b79555c5fe25bc1f436fa&start=25), but I have always used the same device and I had never seen this message.

How can I eliminate the error?

I am afraid that it will reappear again if I leave the raspberry on for a long time on a single image. (I have not yet tried with multiple images if the error appears but I think it is not visible because the screen updates continuously)

38

@caimano, thanks for sharing your results. I checked the link that you shared, and it looks like you can modify /boot/cmdline.txt (for Bullseye and older) or /boot/firmware/cmdline.txt (for Bookworm).

You can try the following. (Make sure to reboot your Pi to see if the issue still persists or not.)

  • Edit cmdline.txt so that quiet will be included, if i’s not yet included. For example,

    console=serial0,115200 console=tty1 root=PARTUUID=e06c52b6-02 rootfstype=ext4 fsck.repair=yes rootwait splash vt.global_cursor_default=0 plymouth.ignore-serial-consoles quiet init=/lib/systemd/systemd net.ifnames=0 cgroup_enable=memory cgroup_memory=1

If you have USB devices attached to any of the USB ports of your Pi, you can try removing it.

39

Hi @nicomiguelino

I have “Bookworm”, I checked the cmdline.txt file and the “quiet” parameter was already present.

So, I restarted the Raspberry and let it run all night on.
Tomorrow morning I will check if the message has reappeared.

PS: I do not have USB peripherals connected to the Raspberry.

Regards

40

In the meantime I would like to ask you another thing about github, on the latest releases page of Anthias, there is a section named WebView-v0.3.2.
What is it ?
Do you suggest to install it?