Direct play on TV takes several minutes to start

[alyflex]

I'm wondering why some of my movies/episodes seems to hang forever/for a very long time (about 3-4 minutes) when I try to start them on my TV, whereas whenever I start them on my PC they basically start instantly.

I thought originally that this was an issue with H265 and that my TV was perhaps incompatible, however the latest episode I just tried to watch on my TV is an H264 file and I can see in the admin dashboard that the file is compatible with my device and that the client is receiving the file without modifications.

I saw this other issue here, but I cannot figure out whether this is a similar issue to the one I'm having or not:
https://forum.jellyfin.org/t-direct-play-issues

I should mention that both my TV and PC are directly cabled into my local network, and thus should have similar transfer speeds.

Any insight would be greatly appreciated

[TheDreadPirate]

Can you share your full jellyfin logs via privatebin.net?

And what are the specs of the server?

[alyflex]

First, thank you for helping.

I'm not really sure how to share things privately using privatebin, but I don't think there is anything private in my logfile anyway, so here is a link to the logfile:
https://privatebin.net/?26bf089d35110843...ZmKooTrYXH
During this time period I was attempting to watch an episode of solo leveling on my TV and on my computer, and it would load instantly on my PC while taking several minutes to start on my TV.
I can generate another cleaner and more isolated log once I am home at my TV again.

Regarding my server specs:
I am running jellyfin on my Truenas scale server version ElectricEel-24.10.2
The server is running jellyfin through docker and has hardware acceleration passthrough.
I have attached the docker compose for the jellyfin container:
jellyfin:
image: linuxserver/jellyfin:10.10.4ubu2404-ls47@sha256:874f2e88485807a6b2f43a93a45f1a2007ba04006e9a96afb80ba15a95eafbb1
container_name: jellyfin
hostname: jellyfin
environment:
- PUID=${PUID}
- PGID=${PGID}
- TZ=${TZ}
volumes:
- ${SSD_DATA_PATH}/jellyfin/configconfig
- ${NAS_DATA_PATH}/mediadata/media # location of media and qbittorrent download folder
- ${SSD_TMP_PATH}/jellyfin/transcodingconfig/cache # location for transcoding
devices:
- /dev/dridev/dri # Video acc
expose:
- ${PORT_JELLYFIN}/tcp # http web ui
ports:
- ${PORT_JELLYFIN}:8096/tcp # http web ui
restart: unless-stopped
networks:
- proxy
labels:
- "traefik.enable=true"
- "traefik.docker.network=proxy"
- "traefik.http.routers.jellyfin.rule=Host(jellyfin.${DOMAIN}) || Host(tv.${DOMAIN})"
- "traefik.http.routers.jellyfin.entrypoints=https"
- "traefik.http.routers.jellyfin.tls=true"
- "traefik.http.services.jellyfin.loadbalancer.server.port=8096"



The actual hardware specs are:
N100 CPU
32 GB RAM
SSD disk for transcoding, configs, but all movies/series exist on HDD
All drivers have lots of free space still.


When I was running these tests yesterday the CPU utilization on my server was very low (which I guess is expected since it was direct play, and the server wasn't doing much else at that time).

Please let me know if I forgot anything.

[TheDreadPirate]

Nothing unusual in the server logs. But if you could share one of the ffmpeg transcode logs, I'd like to see how well your system is performing.

The ffmpeg log is in the same directory as the jellyfin log. The logs that start with FFmpeg.Transcode.

Also, are you using cloudflare tunnels or proxies?

[alyflex]

Here is one transcode log:
https://privatebin.net/?1ac2a71d7261a186...ujxYTwUAxW
and here is another:
https://privatebin.net/?25b9c3e2278d410b...ZEEB8aBPgW

I have traefik setup as a reverse proxy, and I do use tailscale, but everything I did in the above case was done locally on my own internal network.

I am curious to hear whether you think my system is performing okay or not.
One thing I am wondering about is how it can be a server issue if there is no transcoding being done?
Could it be that even though it is supposed to be a direct play my TV still has to do some processing or prebuffering before it can start the episode?
If that is the case then it obviously makes sense that my computer is able to do this much faster than my TV which probably have a pretty pitiful processor, I just wonder why I haven't seen this with everything then.

[TheDreadPirate]

I took a look at your first log and most of the IPs are public IPs. Which tells me the clients aren't local or your local clients are being routed out to the Internet before coming back.

Do you have NAT hairpin enabled on your router?

[alyflex]

Hmm that is strange. I am the only user on this jellyfin and have only used it at home over the last couple of months, so there definitely should not be any public IPs connecting to my jellyfin.
I thought I had successfully set up split horizon dns, but I guess something must be going wrong with that.

I use pihole for my local dns and tailscale dns.
If I try to use

dig jellyfin.mydomain.com

; <<>> DiG 9.18.30-0ubuntu0.24.04.2-Ubuntu <<>> jellyfin.mydomain.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 17236
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 65494
;; QUESTION SECTION:
;jellyfin.mydomain.com. IN A

;; ANSWER SECTION:
jellyfin.mydomain.com. 0 IN A 192.168.0.4

;; Query time: 0 msec
;; SERVER: 127.0.0.53#53(127.0.0.53) (UDP)
;; WHEN: Wed Mar 26 22:19:52 CET 2025
;; MSG SIZE rcvd: 64

Which seems fine to me, but maybe it is my TV that is for some reason not getting it correct.
I will definitely investigate further, thank you very much.

[alyflex]

Interestingly, the jellyfin activity dashboard is also showing that something is not working as it should.

I noticed that it was showing a public IP address as well for my current login when I was going to jellyfin.mydomain.com

Even though I am using the computer where 'dig jellyfin.mydomain.com' is correctly pointing to my local server ip address.
However, when I access the domain through the local port: 192.168.0.4:2296 (which is the port that jellyfin is also exposed on locally, then my ip shows up as:

tue successfully authenticated
today at 10:42 PM
IP address: 192.168.0.174
tue is online from Firefox
today at 10:42 PM
IP address: 192.168.0.174

So... I will report back once I have figured out why things aren't working.