All Activity

The Dashboard web pages where the Now Playing information shows, have websocket connections to the server which have data sent to them whilst there are playback sessions. We discovered that on Windows if the browser tab is suspended - eg when the browser detects it is not in focus and not current active tab, then the feeds from the server buffer up leading to memory usage increase over time. If you do normally have a dashboard tab open, see if closing that browser tab before the machine gets put to sleep makes a difference

Yes you can mix different lineups - so that is one way to get round the problem. I have submitted a request to Gracenote to add the WETADT 26.x set of channels to the Elkton OTA lineups I see from https://www.channelmaster.com/pages/tv-antenna-map-elkton-md-21921 that the channels do not appear for this area. Also for Baltimore this shows them as weak signal channels https://www.channelmaster.com/pages/tv-antenna-map-baltimore-md-21224 These are the channels that i asked gracenote to add - please confirm i have the names correct 26.1 WETA-HD PBS "WETA HD" 26.2 WETA UK PBS Encore "WETA UK" 26.3 KIDS PBS Kids 24/7 "WETA Kids" 26.4 WORLD World Channel "WETA World" 26.5 METRO PBS Encore "WETA Metro"

كمبيوتر ↓ Emby Server ↓ أفلام + مسلسلات ↓ الراوتر / الإنترنت ↓ تلفزيون Samsung ↓ تشغيل الفيلم

I'd like to use a monitor with speakers as a playback device. Are there certain types monitors that are better suited than others for video playback? Thanks

The Nvidia Shield 2019 pro is out of stock everywhere and Nvidia announced a price increase to $300 if or when it becomes available again. Bad news either way. What is an alternative device that runs Android TV, interfaces via HDMI, and supports side-loading and OS customizations like the Shield? It does not need to be as "beefy" as a Shield. Thanks

Current efforts should focus on avoiding redundant work that wastes performance.

Based on the summary it seems very similar if not identical to this

Yes I am using OTA Emby Guide Date - USA-OTA21921. I also add a second guide which did have the missing channels in it (USA-OTA21224). I didn't realize you can use more than one guide.

Hi @Luke, Just wondered if any updates? Bob

That is by design. Otherwise any new icon positions wouldn't be shown and all posters would be stale. I'm planning on making it profile dependant so only affected libraries gets updated and also only posters with relevant information gets updated. The last one might take a while to do.

I close my MacBook down when im not using it, so the laptop is asleep I guess. I generally ALWAYS have an emby tab open when I am using the MacBook. Just on the admin page.

FWIW i had a similar issue of "download screen not fully populating with all normal fields". I had just installed the beta. I dont know what fixed it and In no particular order. 1. enabling photo access (but not targetting a server) removed pages (1000s) of errors (multiple 10s-100s per second) in the ios client log about "no photo access" (ive seen this on previous versions of client as well) 2. removing the device from devices and relogging in the client. one of those fixed the issue for me.

Could it be the suspended dashboard browser tab issue I mentioned above here ? Do you have a browser showing the Emby Server dashboard ? If you do, could it be backgrounded / suspended ? Have a look at the link I provided If it is not this and there is no scanning, i would need to see a debug log with second by second memory usage logging which i can provide you a shell script for

I've discovered a new issue: you cannot save any plugin settings, otherwise all your images will have to be reprocessed.

Thanks @Luke I understand well the process and the parameters needded to add in the server side (preferred language for images, fallback language, and probably an option to have a "strict language code enforcement" that can prevent using fr-CA instead of fr-FR)... I'll keep an eye on the next releases.

And its crashed again. No warning, lasted just 30 minutes. This time it restarted itself however without my automated docker restart process if it looses contact.

I am assuming you are using an OTA Emby Guide Data lineup. Can you please tell me which OTA TV Market you are on ? and also the Guide ID which you can see on the Emby Server Live TV Setup screen. Also please let me have the callsigns for each of the 26.x channels as they appear on the tuner This is an example of an Emby Server Guide Lineup ID

I still deleted the download list, deleted all data from the app, then uninstalled and reinstalled the app, the same.

I have a couple bug fixes marked down to work on.. But waiting for life to slow down before I will have time to work on it. But other than a bug or two, it's pretty much complete.

@Aleas are you planning to rollout some updates any time soon ? Or the project is finished

It used all the ram again - and it is not linear, as it seemed to do it before the 5 minute growth check - [2026-10-05 15:20:02] MEMORY WARNING: Emby is currently drawing 10.05GiB of system RAM. [2026-10-05 15:25:03] MEMORY WARNING: Emby is currently drawing 10.06GiB of system RAM. [2026-10-05 15:30:02] MEMORY WARNING: Emby is currently drawing 10.04GiB of system RAM. [2026-10-05 20:20:02] MEMORY WARNING: Emby is currently drawing 16GiB of system RAM. And then my other script saw it busted, so restarted it : [2026-10-05 20:20:11] ALERT: Emby URL failed health check (Status Code: 000). Restarting container. Just 7 direct play shows. Just normal playing progress in the hour run up to this failure, no scans or anything strange. No errors or anything - 2026-10-05 20:18:44.798 Info PlaystateService-0HNP0B058PKJA:00000012: http/1.1 Response 204 to host87. Time: 28ms. POST http://emby_remote_ip/emby/Sessions/Playing/Progress?X-Emby-Client=Emby for Android&X-Emby-Device-Name=Family room TV&X-Emby-Device-Id=61b83ac49a1b16d9&X-Emby-Client-Version=3.5.55&X-Emby-Token=x_secret57_x&X-Emby-Language=en-au&reqformat=json. 2026-10-05 20:18:44.967 Info PlaystateService-0HNP0B058PKJA:00000013: http/1.1 POST http://emby_remote_ip/emby/Sessions/Playing/Progress?X-Emby-Client=Emby Xbox&X-Emby-Device-Name=DOPYDRAGONXBOX&X-Emby-Device-Id=yNjZdXKrZtaCiJjgbfeEYW8CZGJJ57utGWXx9rZis&X-Emby-Client-Version=2.317.70.0&X-Emby-Token=x_secret61_x&X-Emby-Language=en-gb&reqformat=json. Source Ip: host94, UserAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/150.0.0.0 Safari/537.36 Edg/150.0.0.0 2026-10-05 20:18:44.975 Info PlaystateService-0HNP0B058PKJA:00000013: http/1.1 Response 204 to host94. Time: 8ms. POST http://emby_remote_ip/emby/Sessions/Playing/Progress?X-Emby-Client=Emby Xbox&X-Emby-Device-Name=DOPYDRAGONXBOX&X-Emby-Device-Id=yNjZdXKrZtaCiJjgbfeEYW8CZGJJ57utGWXx9rZis&X-Emby-Client-Version=2.317.70.0&X-Emby-Token=x_secret61_x&X-Emby-Language=en-gb&reqformat=json. 2026-10-05 20:18:45.277 Info PlaystateService-0HNP0B058PKJA:00000014: http/1.1 POST http://emby_remote_ip/emby/Sessions/Playing/Progress?X-Emby-Client=Emby for Android&X-Emby-Device-Name=Smart TV Pro&X-Emby-Device-Id=2b464082bda8a69b&X-Emby-Client-Version=3.5.55&X-Emby-Token=x_secret2_x&X-Emby-Language=en-au&reqformat=json. Source Ip: host2, UserAgent: Mozilla/5.0 (Linux; Android 11; Smart TV Pro Build/RP1A.200622.001; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/153.0.8010.36 Mobile Safari/537.36 2026-10-05 20:18:45.281 Info PlaystateService-0HNP0B058PKJA:00000014: http/1.1 Response 204 to host2. Time: 4ms. POST http://emby_remote_ip/emby/Sessions/Playing/Progress?X-Emby-Client=Emby for Android&X-Emby-Device-Name=Smart TV Pro&X-Emby-Device-Id=2b464082bda8a69b&X-Emby-Client-Version=3.5.55&X-Emby-Token=x_secret2_x&X-Emby-Language=en-au&reqformat=json. Anything I can do here ? Seems like the out of memory issue takes longer to happen now, but it still does it.

Hi, with last version I cannot make any recording of live tv. embyserver.txt

Synopsis generated by AI. Raw logs attached. Setup - Emby Server 4.10.0.40, Ubuntu 24.04 (kernel 6.8), Hyper-V guest, Quadro P2000 (NVENC/NVDEC) - Tuner: HDHomeRun (native HDHomeRun tuner type, stream sharing automatic) - Recording: "NFL Football - Denver Broncos at San Francisco 49ers", channel 862 KCNC-TV, timer 12b26b50ce49aad1077546b3d18c9c95 - Remote viewer: user "Chuck", Emby for Samsung 2.3.6, watching the same channel live (transcoding, low bitrate cap) Summary - 14:00:58 viewer opens KCNC-TV live -> "consumerId added: d635760f..." -> consumer count 1. - 14:23:01 recording timer fires and attaches to the SAME shared stream ("Will record ... for 246.98 minutes"). NOTE: logged as "LiveStream consumerId added: " with an EMPTY consumer id -> consumer count 2. - 14:26 - 15:17 viewer stops/starts repeatedly (failed transcodes on the client side). Every new viewer session logs "consumerId added: <PlaySessionId>" / "count is now 2"; every stop logs "consumerId removed" / "now 1". - 15:17:34 viewer session bb9fc8fb removed -> consumer count 1 (only the recording remains). - 15:18:04 viewer starts NEW session 0581f67e (PlaybackInfo, AutoOpenLiveStream=true) on the same stream. NO "consumerId added" and NO "consumer count is now 2" is logged for this session. - 16:44:59 viewer stops 0581f67e -> "LiveStream consumerId removed: 0581f67e..." -> "consumer count is now 0" -> live stream closed -> "Recording stopped" (~2h22m into a 246.98-minute recording). - No timer delete/cancel request was made. The recording was terminated only because the shared stream was closed. Anomalies 1. The recording registers as a consumer with an empty consumer id. 2. A consumer id that was never added (0581f67e) is "removed" and the count is decremented, reaching 0 while the recording is still active, which closes the stream out from under the recording. Expected: a viewer stopping playback must not close a live stream that an active recording is still consuming. emby-livetv-consumer-count-bug.txt

Rendering the image isn't difficult; the challenge lies in having to redraw it every time—though that issue seems to have been resolved. Now, all that's needed for perfection is to reduce memory consumption. The need to redraw it every time was the biggest problem, but it appears that issue has already been solved.

In reality, relying on disk I/O is indeed the best approach, given that Emby's memory management and its use of SQLite are notorious issues. Since Luke has consistently been reluctant to switch to PostgreSQL and Redis, other plugins really shouldn't consume memory; otherwise, the risk of a system crash becomes extremely high.