LiveKit Community

It successful started <ws://localhost:7880>

but it does not connect to <ws://192.168.0.115:7880> /on my local ip/

You could do `livekit-server --dev --bind 192.168.0.115`

I’m not sure what’s causing the issue here, but you may want to look into using <https://docs.livekit.io/guides/egress/#track-egress|track egress> - it sounds like it does exactly what you’re trying to do

image.png

Heres the screenshot from the connection test.

does the VM have a firewall configured? what's the output from
```$ sudo iptables -L -v -n```

This is the output from the command. Is my docker instance isolated?

yeah if you're using the deploy to vm docs and the docker-compose configuration I believe the services are using 'host' networking so they should be network accessible

you feel confident that the websocket url and the token you provided to the test is correct?

I copied the entire content in the cloud-init file to the user data field when I was initializing the VM in Digital Ocean. I checked the file content, and it did specify using 'host' networking in docker-compose.
The Websocket url I use is <wss://livekit.mydomain.chat>.
The token is generated using the API key and secret from the livekit.yaml file.
The OS of my VM is Ubuntu 22.04 x64

are there any helpful livekit server logs?

The access token is created using livekit-cli with --admin param. Any privilege for the access token should be fine right?
Heres the server logs, getting some 401 error.

yeah that looooks like a auth token issue?

I don't know that you should use the admin param?

your token must have `roomJoin` permission set, and a `room` specified

It disallows joining a room for an admin token

Yes! The connection test established after specifying the room, join room permission and identity param. Thanks team!

Hi
how much CPU/RAM dose livekit needs for 1000 simultaneous conferences?
imagine there are 2 participants in each conferece and participants are only sending voice
I know it depends! but give me some number if you can.
thanks
dose livekit support p2p?!

Hey Prasad, what do you see in your server logs when the connection attempt fails?

I think you may have hit a bug that we've resolved recently in 1.2.1 (TCP/TURN fallback with Go SDK).

Did you enable `allow_tcp_fallback` in the config?

Thanks much, David! :pray: we don't get any `ERROR` messages in the logs, other than `JOIN_TINMEOUT` when this happens:
```2022-09-25T11:45:23.841Z	INFO	livekit	rtc/participant.go:710	closing participant	{"room": "testroom", "roomID": "RM_2ZSSWuL74qxP", "participant": "publisher", "pID": "PA_HuUcwKmAZoFp", "remote": false, "sendLeave": true, "reason": "JOIN_TIMEOUT"}```
we also notice a significant drop in the "online users" when this happens, we are currently at 1.2.0 and we don't have `allow_tcp_fallback` in the configs, will add it and also upgrade to v12.1 and let you know how it goes, thanks again for your help!

hey <@U01KZ61GV5Z> I have updated our cluster to v1.2.1 and have also enabled the `allow_tcp_fallback` in the configs, unfortunately this didn't help to fix the intermittent problem, we still get `could not connect after timeout` message frequently from the following and also the user activity in our app drops significantly when this happens, a restart of the livekit deployment fixes both. we already have `debug` log enabled but surprisingly there are no errors/debug messages when this happens.
```livekit-cli create-room --url <https://livekit.something.com> --api-key XXXXX—api-secret XXXXX --name restroom (&lt;--works always)

livekit-cli join-room --url <https://livekit.something.com> --api-key XXXXX --api-secret XXXXX --room testroom --identity publisher --publish /tmp/livekit_audio.ogg (&lt;--fails intermittently)
could not connect after timeout```

can you upload the full logs from the server before + after this happens?

Sure, I will share the logs when this happens next time, thanks! :pray:

Hey <@U01KZ61GV5Z> please find the working logs at <https://gist.github.com/prasadkris/2aaa6e8d02b210268ceb39c6e70cafaa>  and the failure logs at <https://gist.github.com/prasadkris/a416121b003a740a7b3f32b48f5aaba1>. let me know if you need any further details, thank you! :pray:

hey <@U01KZ61GV5Z> were you able to find out anything from the logs, kindly let me know if you need any further details, thanks! :pray:

Hey, I haven't had a chance to look yet, still on my list.

Hello <@U01KZ61GV5Z> I always get the following when it fails.. any idea what might be the issue here, thank you! :pray:
```2022-09-29T05:21:05.419Z	INFO	livekit	service/roommanager.go:254	starting RTC session	{"room": "prasadtest3", "nodeID": "ND_yKNKtfJ8", "participant": "publisher", "sdk": "GO", "sdkVersion": "0.10.5", "protocol": 8}
2022-09-29T05:21:05.420Z	INFO	livekit	rtc/room.go:287	new participant joined	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "pID": "PA_HdEcTCqpkDDY", "participant": "publisher", "protocol": 8, "options": {"AutoSubscribe":true}}
2022-09-29T05:21:05.421Z	DEBUG	livekit	rtc/participant.go:1190	updating participant state	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "state": "JOINED"}
2022-09-29T05:21:05.421Z	DEBUG	livekit	rtc/room.go:236	participant state changed	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "state": "JOINED", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "oldState": "JOINING"}
2022-09-29T05:21:05.421Z	DEBUG	livekit	rtc/participant.go:779	SetMigrateState	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "state": "MIGRATE_STATE_COMPLETE"}
2022-09-29T05:22:05.422Z	INFO	livekit	rtc/room.go:393	closing participant for removal	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "pID": "PA_HdEcTCqpkDDY", "participant": "publisher"}
2022-09-29T05:22:05.422Z	INFO	livekit	rtc/participant.go:640	try close participant	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "sendLeave": true, "reason": "JOIN_TIMEOUT"}
2022-09-29T05:22:05.422Z	DEBUG	livekit	rtc/participant.go:1190	updating participant state	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "state": "DISCONNECTED"}
2022-09-29T05:22:05.423Z	INFO	livekit	rtc/participant_signal.go:199	closing signal connection	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false}
2022-09-29T05:22:05.424Z	INFO	livekit	service/roommanager.go:433	RTC session finishing	{"participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz"}
2022-09-29T05:22:05.424Z	INFO	livekit	rtc/transport.go:1292	leaving events processor	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "transport": "PUBLISHER"}
2022-09-29T05:22:05.424Z	INFO	livekit	rtc/transport.go:486	peer connection state change	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz", "participant": "publisher", "pID": "PA_HdEcTCqpkDDY", "remote": false, "transport": "PUBLISHER", "state": "closed"}
2022-09-29T05:22:48.349Z	INFO	livekit	rtc/room.go:546	closing room	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz"}
2022-09-29T05:22:48.350Z	INFO	livekit	service/roommanager.go:107	deleting room state	{"room": "prasadtest3"}
2022-09-29T05:22:48.350Z	INFO	livekit	service/roommanager.go:401	room closed	{"room": "prasadtest3", "roomID": "RM_hXm4jXmyAmqz"}```

hey <@U01KZ61GV5Z> Have you got a chance to look into this?, thanks!

hey Prasad, sorry I have not been able to. I see you've filed an issue. that's great. We'll follow up with you there

Thanks much <@U01KZ61GV5Z> :pray: No problem, I understand that you guys are really busy as you need to help 1000+ members here, I filed an issue as it is easy to collaborate in GH :slightly_smiling_face:, Thanks again for all your help and having this community channel! :+1:

Hello guys, sorry for my bad english... :wink:
I just set up livekit on my VM where there is also foundryVTT server; i deployed livekit with your instructions but i saw that livekit conflict with foundry's nginx proxy server on port 443 because livekit run automatically caddy that listens 443 TCP port.

so I commented caddy's part on docker startup services script and I configured nginx to proxy and server livekit (ports routing and certificates), I opened ports on firewall but I can't get the connection tester to work....
some ideas?

Schermata a 2022-09-26 13-22-32.png

This is my situation.... with both caddy enabled (and nginx disabled) and with nginx enabled (and caddy disabled)
my turn domain is working with his certificate and point to my livekitdomain...
helppp :wink:

just to follow up on this, I think it's because of mix of archs (im on a Mac M1)

Were you running a binary compiled for intel? That's interesting to know. I've never seen these errors before.

<@U01KZ61GV5Z> thanks for your response. I'd built the container from source using my M1

I've got the deployment working (partially) in ECS using our CI/CD pipelines to build the image. The deployment is up and healthy but I'm trying to determine why I can't connect to it using the example app

turn up and running, but no way to have ok on test page....

Schermata a 2022-09-26 16-43-00.png

Hi <@U043XULAS4C>, could you please thread your messages? It makes it difficult for others to follow the conversation and help with your reported issue(s). :pray::skin-tone-4: 

docker run --rm -it -v$PWD:/output livekit/generate

i have same behavior with both configuration: mine with nginx and caddy's automatically generated;
this my netstat output "tcp        0      0 0.0.0.0:5349            0.0.0.0:*               LISTEN      6494/livekit-server" so turn is listening

we do not handle nginx based deploys. It should work assuming TLS termination is configured correctly

i have same behavior with caddy that has been configured automatically....

when i call my livekit webpage i see "OK", if i call turn's webpage what should i see?

TURN/TLS wouldn't serve anything to your browser

firewall ports are open, livekit is listening on port 5349, caddy is listening on port 443.... what could be the problem?

ok guys nobody wants to help me; i tried absolutely everything, i don't know how to fix, i tried in every way but i failed.
livekit does not have a discord server, no tutorials,  no resources; the only one is <http://livekit.io|livekit.io> and you (here) so I have no other solutions.
Thanks for your help.
See you.

<@U043XULAS4C> first off, this (Slack) _is_ the equivalent of a Discord server. Second, our docs on <https://docs.livekit.io/deploy/|deployment> has worked for many.

We try our best to help everyone that has questions or needs advice, but it does require some patience, we usually catch up on community messages in the evenings and weekends.

I asked you for help because I read ALL your guides and documentation on deployment.
I spent hours trying.
It's a strange situation, I'm not new to Linux, I know how it works but I can't solve this problem so I need someone who knows the livekit application.
I'll wait for someone to help me

as Russ mentioned, this is a community channel where it's not guaranteed someone will know how to solve your specific problem.

we are making LiveKit open for all to use, but there isn't enough time in a day to help everyone deploy it. What I can say on this is folks rarely report issues using the generated deployment on a simple provider like Digital Ocean.

I would suggest trying the above combination first

<https://pastebin.com/05Vbm2qB>
This my livekit-server verbose log.
I don't see any errors but your test keeps giving me error on the turn connection.

Schermata a 2022-09-29 16-05-00.png

Schermata a 2022-09-29 16-09-44.png

turn and caddy are listening..
Some idea?

The only suspect string ithat i can see is:
1. `2022-09-29T15:57:40.860+0200    INFO    livekit rtc*/*participant.go:640  try close participant   *{*"room": "stark-tower", "roomID": "RM_8VdJUKwkgs5m", "participant": "tony_stark", "pID": "PA_NyjMABaaTV8m", "remote": *false*, "sendLeave": *true*, "reason": "DUPLICATE_IDENTITY"*}*`
maybe turn-server is ok and there is a token problem?

yeah, I think you are right. I've just confirmed that with a fresh setup, the TURN test isn't passing, but I believe it's a test issue.

TURN/TLS should be working correctly.

ok, confirmed TURN/TLS is working properly even though test is indicating otherwise. we'll look into improving the test itself, but your setup seems to be working just fine

Thanks for pushing on this <@U043XULAS4C>. :pray::skin-tone-4: 

please let me know (ping me) when the tester will ok :wink:

hi guys, I'm back with the same problem.... livekit tester can't verify my turn, is it still the old problem or is it my problem this time?

<@U01KZ61GV5Z> <@U01KZ0SCDKL> maybe livekit-tester has been fixed and now works fine in detecting turns?

Livekit (and Turn) server started without problems, TLS endpoint is in another PC that run as a proxy server, so unencrypted traffic is redirected to livekit/turn server PC on port 5349

Schermata a 2023-10-11 23-20-05.png

your TURN server isn't setup correctly. likely due to how TLS is configured. Note that you have `external_tls` set to true

this is correct, my certificates are in another server, a reverse proxy with nginx, so turn incoming traffic is unencrypted on port 5349, for this scenario my conf still uncorrect?

the config is fine in that case, but it's most likely your nginx server is configured incorrectly.

I don't know what to try anymore, could you please tell me which ports nginx should listen to and on which ports of the PC with livekit it should proxy?

Currently, the router is forwarding incoming TCP port 5349 to the Nginx server, which decrypts the data and forwards the traffic to TCP port 5349 on the machine with the livekit server, using the configuration I have already sent you.

given the complexity of deployment environments.. we really aren't in a position of troubleshooting each install with users.

what I can say is most folks have no problems following the automatic Caddy setup that we've provided <https://docs.livekit.io/realtime/self-hosting/vm/|here>.

Perhaps you could implement the missing functionality and submit it back as a PR to the repo maintainer?

<@U032TUAL4KT> implemented egress functionality into his PHP SDK, he may have some tips. :blush: 

Its a bit hard for me to help without knowing much about Python, but which part are you having issue with? <@U03S4JSQPFC>

can you be more specific? which SDK/platform are you referring to. is there a bug filed?

Im using JS sdk. In chromium based browsers there are restriction for autoplaying audio tags. And i see that error sometimes in my app that audio can be played only using user gestures

Did you check out <https://github.com/livekit/client-sdk-js/#audio-playback|Audio playback docs> ?

Nope. I had missed this. Sorry for taking your time :pray:

This is working. But whenever a new participant connects, it asks for user gesture. is this how it works ?

it shouldn't do that.. can you turn on debug logging `setLogLevel('debug')` and post the client logs that you see when new participants join?

chunk-vendors.e0e9125d.js:316 room event {event: 'participantConnected', args: Array(1)}args: [Ba]event: "participantConnected"[[Prototype]]: Object
chunk-vendors.e0e9125d.js:316 room event {event: 'trackPublished', args: Array(2)}
chunk-vendors.e0e9125d.js:316 room event {event: 'connectionQualityChanged', args: Array(2)}
chunk-vendors.e0e9125d.js:316 connection quality changed <mailto:karthiksirsikar@teslon.io|karthiksirsikar@teslon.io>-3 excellent
chunk-vendors.e0e9125d.js:316 received server offer {RTCSdpType: 'offer', signalingState: 'stable'}
chunk-vendors.e0e9125d.js:316 room event {event: 'trackSubscriptionPermissionChanged', args: Array(3)}
chunk-vendors.e0e9125d.js:316 room event {event: 'trackSubscribed', args: Array(3)}
chunk-vendors.e0e9125d.js:316 subscribed to track TR_AMtMsc8K8Hph6E <mailto:karthiksirsikar@teslon.io|karthiksirsikar@teslon.io>-3
app.a3c7efa6.js:1 trackSubscribed Ba {_events: {…}, _eventsCount: 12, _maxListeners: undefined, audioLevel: 0, isSpeaking: false, …} Ha {_events: {…}, _eventsCount: 7, _maxListeners: undefined, metadataMuted: false, handleMuted: ƒ, …}
chunk-vendors.e0e9125d.js:316 could not playback audio DOMException: play() can only be initiated by a user gesture.
Ns.handleAudioPlaybackFailed @ chunk-vendors.e0e9125d.js:316
Or.emit @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
Promise.catch (async)
attach @ chunk-vendors.e0e9125d.js:316
attach @ chunk-vendors.e0e9125d.js:316
mounted @ app.a3c7efa6.js:1
In @ chunk-vendors.e0e9125d.js:63
Ni @ chunk-vendors.e0e9125d.js:63
insert @ chunk-vendors.e0e9125d.js:63
O @ chunk-vendors.e0e9125d.js:63
(anonymous) @ chunk-vendors.e0e9125d.js:63
Oi.e._update @ chunk-vendors.e0e9125d.js:63
i @ chunk-vendors.e0e9125d.js:63
e.get @ chunk-vendors.e0e9125d.js:63
e.run @ chunk-vendors.e0e9125d.js:63
Gi @ chunk-vendors.e0e9125d.js:63
(anonymous) @ chunk-vendors.e0e9125d.js:63
Vn @ chunk-vendors.e0e9125d.js:63
Promise.then (async)
Hn @ chunk-vendors.e0e9125d.js:63
Kn @ chunk-vendors.e0e9125d.js:63
Xi @ chunk-vendors.e0e9125d.js:63
e.update @ chunk-vendors.e0e9125d.js:63
e.notify @ chunk-vendors.e0e9125d.js:63
set @ chunk-vendors.e0e9125d.js:63
(anonymous) @ app.a3c7efa6.js:1
u @ chunk-vendors.e0e9125d.js:216
(anonymous) @ chunk-vendors.e0e9125d.js:216
(anonymous) @ chunk-vendors.e0e9125d.js:216
a @ chunk-vendors.e0e9125d.js:79
s @ chunk-vendors.e0e9125d.js:79
(anonymous) @ chunk-vendors.e0e9125d.js:79
t @ chunk-vendors.e0e9125d.js:121
(anonymous) @ chunk-vendors.e0e9125d.js:79
t @ app.a3c7efa6.js:1
(anonymous) @ app.a3c7efa6.js:1
r.emitEvent @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
Or.emit @ chunk-vendors.e0e9125d.js:316
emit @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
Or.emit @ chunk-vendors.e0e9125d.js:316
emit @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
Or.emit @ chunk-vendors.e0e9125d.js:316
setTrack @ chunk-vendors.e0e9125d.js:316
addSubscribedMediaTrack @ chunk-vendors.e0e9125d.js:316
onTrackAdded @ chunk-vendors.e0e9125d.js:316
(anonymous) @ chunk-vendors.e0e9125d.js:316
Or.emit @ chunk-vendors.e0e9125d.js:316
cr.subscriber.pc.ontrack @ chunk-vendors.e0e9125d.js:316
a @ chunk-vendors.e0e9125d.js:316
a @ chunk-vendors.e0e9125d.js:316
Show 23 more frames
chunk-vendors.e0e9125d.js:316 room event {event: 'audioPlaybackChanged', args: Array(1)}args: Array(1)0: falselength: 1[[Prototype]]: Array(0)event: "audioPlaybackChanged"[[Prototype]]: Object

this is when 3rd partcpant joined. i had shown a button on ui when the 2nd participant joined and called `startAudio()` after the click.

and i noticed one more thing. When i first called `startAudio()` the `AudioPlaybackStatusChanged` event got fired but the `room.canPlayAudio` was still false. and when i call `startAudio()` again `room.canPlayAudio` became true

please include the full logs from when the first participant joined. Uploading to a Github gist (or something similar) would be helpful.

<https://gist.github.com/karthikSirsikar/f45c82319b3938d62d1a76b1becfc10b>

thanks, it does appear to be firing `audioPlaybackChanged` multiple times on this browser/OS, could you file an GH issue in client-sdk-js repo? Please include repro steps (or a snippet of code that could reproduce the issue). Thanks!

<https://github.com/livekit/client-sdk-js/issues/447>

can you specify which app you are using to test? is this our example app?

Please indicate that in the issue. Having no reproduction steps will make it unlikely for the engineers to know what to do with the ticket.

Have edited the issue <https://github.com/livekit/client-sdk-js/issues/447>

not able to reproduce the issue you'd reported, but we are seeing some other strange behaviors on MS Edge mobile.

<@U02EU2ZPTQU> Any update on issue #447 ?

Hi Karthik, just to set expectations on timing. This is an open source community, and we try to help everyone to the best of our abilities.

However, this isn't a commercial support channel where there are guarantees around resolution or when we will be able to get to it. If the issue is urgent, I would suggest trying to play with the code base and see if you can make progress on it. (one benefit of open source!)

we'd welcome the help and would be glad to accept PRs!

Where can i find the docs for `expWebAudioMix` feature to know what it does actually

it's implemented in <https://github.com/livekit/client-sdk-js/pull/446|this PR>

Im not good at typescript and i've never worked on an open source project so i wouldn't be able to do a PR now. But i'll look forward to do it anyway in the future as it is an interesting project.

For time being i've added a crude fix in my wrapper app where i render some 50 audio tags on the start itself and attach tracks as and when a participant joins. By doing this no audio tags are created later on and browsers also don't crib about it

ah ha, that's a great workaround. thank you for sharing!

1664250178791.png

try running without root. Root user may have restrictions from the container when writing to the host OS

hi!
you can just call `await room.localParticipant.setCameraEnabled(false);` to disable the camera

thanks for response <@U02EU2ZPTQU>,
and is there any way to know that a participant has disabled the camera?

sure, have a look at the react example here <https://docs.livekit.io/guides/room/events/#declarative-ui>
additionally there’s also `localParticipant.isCameraEnabled`

What you are describing sounds like flakey networking to the client and/or how your server is deployed.

We use long-running websocket connections that certain load balancers would terminate after some time

we don't have any load balancer or firewall
we can not use something like bosh (http long polling) in livekit?
we don't have good internet in our country. livekit clinets can not configure to reconnect after any disconnections?

Clients will reconnect when they detect interruptions. it'll only give up after it has exhausted all connection attempts and still unable to reconnect.

If you'd like to keep retrying, please listen to the Room disconnected event and reconnect to the server

thanks you so much
in our experience we both was in room but one of use can not hear another (i am not 100% sure that the other one can hear or not) in this case disconnected event raises?

Are you running v1.4.1? This should be fixed

root@mr01:/opt# docker inspect bc3 | grep image
                "org.opencontainers.image.created": "2022-09-15T18:11:32.158Z",
                "org.opencontainers.image.description": "Export and record WebRTC sessions and tracks",
                "org.opencontainers.image.licenses": "Apache-2.0",
                "org.opencontainers.image.revision": "adae4d308d7a9179c7c6cdf755918b244180edfa",
                "org.opencontainers.image.source": "<https://github.com/livekit/egress>",
                "org.opencontainers.image.title": "egress",
                "org.opencontainers.image.url": "<https://github.com/livekit/egress>",
                "org.opencontainers.image.version": "v1.4.1"
root@mr01:/opt#

Could it be related to GST_PLUGIN_SCANNER?

no, the error is
```gst_matroska_mux_video_pad_setcaps (): /GstPipeline:pipeline/GstBin:input: \nCaps changes are not supported by Matroska```

egressrec01 | (egress:21): GStreamer-WARNING **: 22:37:22.873: External plugin loader failed. This most likely means that the plugin loader helper binary was not found or could not be run. You might need to set the GST_PLUGIN_SCANNER environment variable if your setup is unusual. This should normally not be required though.

Do you have any suggestions for a solution?

the track changed from 360x640 to 540x960, which matroska mux does not allow

I’m currently building a tool to better reproduce and test this case, and we should have a better fix in 1.4.2

for now, downgrading to v1.2.0 will fix it but you will receive .ivf files instead of .webm files

yes, if you need this to work immediately

otherwise, I will be releasing 1.4.2 sometime this week once I’m confident in the new fix

Yes, it continues to record with v120 without giving an error. thank you for your help.

What are your suggestions for stable operation in long (1 hour) broadcasts?

it probably is taking a long time to finalize/upload the file

0 byte file is a sign that the process was killed before it was able to complete

make sure it’s being stopped with StopEgress and that the docker container isn’t being killed

Just pushed 1.4.2 - I was able to reproduce in 1.4.1 but not 1.4.2, but let me know if you see it again

Thank you for your help. If I try v1.4.2 I will let you know. good work

hi jibon!
thanks for the report, changing the default value was unintentional, will release a fix today with the correct default value! 

we haven't ran any tests on Fargate and cannot speak to its compatibility there.

if you've disabled external IP resolution,  you'd have to set --node-ip to the correct public IP for each instance of livekit. With WebRTC, the server *must* know how to advertise itself to clients.

Thanks <@U01KZ61GV5Z>, after some investigation it appears to me that it's entirely around the wrong ip resolution, and as a result I'm getting the timeout

Looks like your livekit server isn't running. How are you deploying livekit? Docker?
From browser can you access <http://localhost:7880/>?

i use livekit-cli

when i use createToken function, its right.

but listRooms function not working TT

this is my createToken function

```<http://app.post|app.post>("/api/room/create", async (req, res) =&gt; {
  const at = new AccessToken('devkey', 'secret', {
    identity: participantName,
  });
  at.addGrant({ roomJoin: true,roomList:true, room: roomName });

  const token = at.toJwt();
  console.log('access token', token);
  return res.status(200).json({result: true});
});```

Are you running the server?, livekit-cli isn't a server

i use this code for running server
``` livekit-server --dev```

It might be a memory issue, v1.2.0 waits until completion to write the file in order to put the headers at the beginning (this was changed in v1.4.2). If your recordings will always be long, another option would be to use segments, which will upload an HLS/m3u8 playlist of 6s segments

I couldn't find a sample code related to HLS. Where should I look?

On the other hand, is there anything you can say about IVF-type recording? It makes a lot of sense to connect to the room as a user and record the broadcast. How can I handle this fast recording (or frame drop). I also looked at galene's code. it is coded in a very similar way.

HLS: <https://github.com/livekit/protocol/blob/main/livekit_egress.proto#L88> replace `TrackCompositeEgressRequest_File` with `TrackCompositeEgressRequest_Segments`

we used to use the pion ivfwriter from our sdk (removed here: <https://github.com/livekit/egress/commit/a1b7e24a9ab42542676f427e9433975ca9c2c921>), however there were issues with sync/playback speed, and things like muting

here’s a better link for how ivf worked <https://github.com/livekit/egress/blob/0d8ddd35d523595a25ef52d449df622cabdca9f9/pkg/pipeline/source/filewriter.go>

<https://docs.livekit.io/guides/egress/room-composite/#stream-to-hls|HLS docs>

Actually, if there is no sound, you should use Track egress. it's simpler and much less intensive computationally.

I am currently using "Track egress" and I have problems with long time (1 hour) recordings. I think if you test this you will find the same situation.

That's why I found IVF healthier and more stable.

But I will also try HLS.

I think LiveKit has made a very good choice to record rooms with headless-chrome unlike other media servers. But I think this issue needs to be analyzed in more depth. Bigbluebutton uses the same method. I haven't encountered such a problem with BBB.

Thank you again for your help.

HLS will not work with track egress as HLS requires transcoding