Artworks not showing after 8.4.0 Update
amaduain opened this issue · comments
This is one of my favourites addons! but the artwork is not working anymore.
I updated to the latest version and I noticed that the artwork stopped working., I cannot see any cover for albums, searches or songs, after some investigation seems that the URLs have changed and now it is not able to show the pictures anymore
When everything worked the URL link of an artwork was this:
"https://www.mysqueezebox.com/public/imageproxy?u=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fq1Vp ipwOxLE%2Fmaxresdefault.jpg"
Same album on the newest release 8.4.0 is like this:
"http://192.168.20.4:8123/imageproxy/https%3A%2F%2Fi.ytimg.com%2Fvi%2Fq1VpipwOxLE%2Fmax resdefault.jpg/image_300x300_f.jpg"
Notice that the URL is for my local home assistant (not real IP) on port 8123 and that URL is not working at all :-(
Any hints?
hmmm this via ingress? What theme are you using might just be worth a quick rescan from scratch also. If it persists please log a issue, but in all my testing all this is ok I don’t have lms configured to use mysb for artwork anyway. This might be related to 8.4.0 deprecating use of mysqueezebox.com where possible.
Take it it all looks ok if you goto http://192.168.20.4:9000
You are right, I'm trying with other example, the URL passed to the browser is this:
http://192.168.20.4:8123/imageproxy/https%3A%2F%2Fi.scdn.co%2Fimage%2Fab67616d0000b2730fc3f45ec3c0037a882b1235/image.png
and it is not working:
But if I change the URL to port 9000 it is working:
So it is a matter on do the proper redirect on the container
This is happening on the albums for the apps, Spotty, Youtube, etc, So I guess it uses dinamically the urls for the results, so the scan is not going to work in this case.
Is it configured as your default skin in LMS.
The good news is on my test system it all seems ok, which probably means this is totally fixable/repeatable if we can get the right info to me and we should be able to work round it in the meantime.
So what are you settings under
server->interface->performance->Artwork resizing
server->interface->Web Interface
Is it configured as your default skin in LMS.
The good news is on my test system it all seems ok, which probably means this is totally fixable/repeatable if we can get the right info to me and we should be able to work round it in the meantime.
So what are you settings under server->interface->performance->Artwork resizing server->interface->Web Interface
Thanks a lot for the help, I've been trying, and seems there is something not right with the material skin.
I did reinstall from scratch, deleting the LMS folder and so on, configured only the youtube app and did a search:
As you can see, the artwork is loaded properly
If I install the material skin and change to it, all the artwork is gone:
LMS->server->interface->Web Interface
Should be Default,
HA-Addons-LMS-Skin to:-
material
server->interface->performance->Artwork resizing
separate helper process.
Clear cache restart browser.
set server->interface->performance->Artwork resizing to separate process by default on new installs.
@pssc - I am unfortunately also having this problem, even when using the settings above. The direct URL on port 9000 works, but ingress has incorrect/incomplete URLs for album art resulting in 404s.
I have my settings as follows, as suggested above:
- Within LMS: Settings > Interface > Web Interface: Default
- Within LMS: Settings > Advanced > Performance > Artwork resizing:
Use Logitech Media Server resizing helper to resize artwork
- In the HA add-on settings: Configuration > Skin:
material
If it's helpful, I also use the NGINX Home Assistant SSL proxy
add-on for HTTPS access to HA.
Can you give an example of a broken url?
It seems all to be happy in my test config.
Things todo:-
Check LMS config
Check addon config and restart
Clear all caches and force reload pages in we browsers
Thanks for the quick reply. Tried those (restart the add-on, different machines, browsers, private browsing), with same results unfortunately.
I can access LMS directly at http://ha.hostname:9000
, and if I do, the album art URLs look like this:
http://ha.hostname:9000/imageproxy/http%3A%2F%2Fresources.tidal.com%2Fimages%2Fdaee3088%2Fdd9f%2F4bbd%2F8a88%2F36767eabccde%2F1080x1080.jpg/image_100x100_o
If I access LMS through ingress (https://ha.hostname/hassio/ingress/7222e6e9_lms
), then the album art URLs look like this:
https://ha.hostname/imageproxy/http%3A%2F%2Fresources.tidal.com%2Fimages%2F3948376a%2Fcc40%2F40b4%2Fa97d%2F429f13d61973%2F1080x1080.jpg/image.jpg
That doesn't seem right, but I don't know what format the URLs are supposed to take in order to pass them through to LMS.
Does it happen for local or is it Tidal only?
Does it work when not going through the nginx https proxy?
Does it happen for local or is it Tidal only?
It's not just Tidal, as BBC Sounds also lacks any album art, but local albums work fine. Those URLs look like this:
https://ha.hostname/api/hassio_ingress/sqblct1HWR1dQALj9NgE2NaiL5xaFFO5vdZNyQLlFpY/music/83ceac23/cover_300x300_f
Does it work when not going through the nginx https proxy?
It does not work with Tidal or BBC Sounds even if I access HA directly at http://ha.hostname:8123
.
Ah yes ok each image url is generated by the plugin code, I will prod BBC Sounds into life on the test system and see what I can find.
Thank you!
Hi @pssc - just wanted to see if you were able to reproduce this, or if there's anything I can do to help do so. Thanks!
@baf I can replicate artwork missing from plugins but, Ii can't find a fix at present
@pssc, understood, thanks. I don't have much experience to offer here but will poke around when I get a chance.
I'm also having the same problem.
@pssc - I have a potential fix.
Change this line to sub_filter_types *;
I'm no nginx expert, so I'm sure it's ham-fisted and could be refined, but wanted to get your thoughts. It does display all album artwork via HA ingress using the Material skin.
That's all right @baf I am not either any suggestions/possible fixes welcome, that implies that there is a content type missing. Given i don't see this if you have chrome you can look at the content types on the network inspection page/tab, looking at it cold though I hope text/html is implied. What OS/Setup are you on for the addon?
Using the dev tools as you suggested, it looks like the relative URLs are part of a JSON response from LMS, so adding application/json
to the same line that I referenced above does the trick (the complete line: sub_filter_types text/plain text/css application/javascript application/x-javascript text/javascript application/json;
). Feels like there should be a better way still, but I haven't come up with it yet.
My setup is Home Assistant OS 13.1, currently with HA version 2024.8.3. I am also using an HTTPS reverse proxy with it and 99% of the time that's how I access HA, so that may complicate matters somewhat as well, although it doesn't cause problems for the rest of my HA setup.
7a85b43 may fix this is anyone able to pull down the next branch locally to try. @baf no I think this may be the only way, are you using most local content or something else? As I was unable to replicate issue, I suspect material is displaying base urls mostly from plugins direct out of the json feed. Thanks for your help on this @baf all really useful info.
Sorry, I should have mentioned that - I use very little local content. The majority of my usage is via the Tidal plugin (https://github.com/michaelherger/lms-plugin-tidal), with some BBC Sounds (https://github.com/expectingtofly/LMS_BBC_Sounds_Plugin) as well.
@baf no need its normal for you. But now with this nugget, I think I understand why I wasn't able to replicate. I will go play with the BBC plugin. I suspect your local content may be ok?
@pssc thanks for the update, I'm using mostly spotify and youtube, those are still not working for me, I stopped the add-on, edited the:
/usr/share/hassio/addons/git/fc57b866/lms/rootfs-ingress/etc/nginx/servers/infress.conf
, but everything stays the same, I don't know if the changes I'm doing on the file are being refreshed on the add-on or not.
@amaduain you will need nginx to reread the config kill -HUP <insert pid of nginx master proc here>
find with ps -axf and do a full refresh in the browser
I can confirm that the changes fixes the BBC plugin for me, I need to finish up a few things in flight around nostats and mmserver toggles and I will push out a new release.
Thanks much!