LiveKit Community

anybody have experiences similar issues on M1?

Hi Timotej, what OS/docker version are you running? that seems really strange.

it seems like something got mixed up on the input.. and it's trying to run the image `for` instead of `livekit/generate`

I have tried reinstalling it and the command `sudo docker run hello-world` works I tried pulling the livekit/generate and it also works

could the issue be that I am running it in an empty folder?

no, it should be ok to run on an empty folder.. the script should be generating configs into the current dir.

you probably have to put quotes around $PWD:
`docker run --rm -it -v"$PWD":/output livekit/generate`

Looks like a crash that would need to be tracked down in gstreamer. Are you using one of our egress containers? What version?

I use the Egress container and update to the latest version few days ago.

Ok. Would it be possible to get the egress start request you are using, and some more verbose debug? You could get this by setting the log level to debug in the egress conf and add the `-e GST_DEBUG=3 command line option to the docker command line`. 

There are debug logs .
```
2022-12-30T03:47:54.875Z	INFO	egress	service/service.go:153	request accepted	{"nodeID": "NE_ug96fQH8tez9", "egressID": "EG_DUGVkTesaSLG", "requestID": "RPC_HRuDFUDF7wqA", "senderID": "ND_ElfvPvE3"}
2022-12-30T03:47:54.907Z	INFO	egress	redis/redis.go:82	connecting to redis	{"nodeID": "NE_ug96fQH8tez9", "handlerID": "EGH_cbZtJrgjtDQs", "egressID": "EG_DUGVkTesaSLG", "simple": true, "addr": "127.0.0.1:6379"}
0:00:00.003690750    25 0x7fe0d4018290 ERROR     GST_PLUGIN_LOADING gstpluginloader.c:442:gst_plugin_loader_try_helper: Spawning gst-plugin-scanner helper failed: Failed to close file descriptor for child process (Operation not permitted)

(egress:25): GStreamer-WARNING **: 03:47:54.912: 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.
0:00:00.074762444    25 0x7fe0d4018290 WARN                 nvcodec gstcudaloader.c:139:gst_cuda_load_library: Could not open library libcuda.so.1, libcuda.so.1: cannot open shared object file: No such file or directory
0:00:00.074794689    25 0x7fe0d4018290 WARN                 nvcodec plugin.c:73:plugin_init: Failed to load cuda library
0:00:00.433007846    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x120000: 'AVR (Audio Visual Research)' is not mapped
0:00:00.433046018    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x180000: 'CAF (Apple Core Audio File)' is not mapped
0:00:00.433059706    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x100000: 'HTK (HMM Tool Kit)' is not mapped
0:00:00.433081592    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0xc0000: 'MAT4 (GNU Octave 2.0 / Matlab 4.2)' is not mapped
0:00:00.433088414    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0xd0000: 'MAT5 (GNU Octave 2.1 / Matlab 5.0)' is not mapped
0:00:00.433094249    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x210000: 'MPC (Akai MPC 2k)' is not mapped
0:00:00.433100843    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0xe0000: 'PVF (Portable Voice Format)' is not mapped
0:00:00.433107752    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x160000: 'SD2 (Sound Designer II)' is not mapped
0:00:00.433116297    25 0x7fe0d4018290 WARN                 default gstsfelement.c:97:gst_sf_create_audio_template_caps: format 0x190000: 'WVE (Psion Series 3)' is not mapped
2022-12-30T03:47:56.960Z	INFO	egress	service/handler.go:133	egress updated	{"nodeID": "NE_ug96fQH8tez9", "handlerID": "EGH_cbZtJrgjtDQs", "egressID": "EG_DUGVkTesaSLG", "egressID": "EG_DUGVkTesaSLG", "request_type": "room_composite", "output_type": "segments", "status": "EGRESS_ENDING"}
0:00:02.054936874    25 0x7fe0dc02e6a0 WARN              aggregator gstaggregator.c:2101:gst_aggregator_query_latency_unlocked:&lt;muxer&gt; Latency query failed
0:00:02.055637455    25 0x7fe0dc02ed20 FIXME                default gstutils.c:4025:gst_pad_create_stream_id_internal:&lt;pulsesrc0:src&gt; Creating random stream-id, consider implementing a deterministic way of creating a stream-id
0:00:02.056145697    25 0x7fe0dc02e580 FIXME                default gstutils.c:4025:gst_pad_create_stream_id_internal:&lt;ximagesrc0:src&gt; Creating random stream-id, consider implementing a deterministic way of creating a stream-id
0:00:02.058775899    25 0x7fe0dc02ecc0 WARN         audio-resampler audio-resampler.c:274:convert_taps_gint16_c: can't find exact taps
**
ERROR:../gst/multifile/gstsplitmuxsink.c:2691:check_completed_gop: assertion failed: (gop != NULL)
Bail out! ERROR:../gst/multifile/gstsplitmuxsink.c:2691:check_completed_gop: assertion failed: (gop != NULL)
2022-12-30T03:47:57.245Z	ERROR	egress	service/manager.go:114	process failed	{"nodeID": "NE_ug96fQH8tez9", "error": "signal: aborted (core dumped)"}
<http://github.com/livekit/egress/pkg/service.(*ProcessManager).awaitCleanup|github.com/livekit/egress/pkg/service.(*ProcessManager).awaitCleanup>
	/workspace/pkg/service/manager.go:114
2022-12-30T03:47:57.245Z	INFO	egress	service/service.go:88	shutting down	{"nodeID": "NE_ug96fQH8tez9"}```

Can you share the request that was made too?

The failure looks really strange as it's attempting to load a hardware encoder. We do not support hardware encoders in Egress today. Which environment are you running this?

```WARN                 nvcodec gstcudaloader.c:139:gst_cuda_load_library: Could not open library libcuda.so.1, libcuda.so.1: cannot open shared object file: No such file or directory```

Also setting the log level to debug in the egress conf would help. 

```Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits virtual
CPU(s):                          4
On-line CPU(s) list:             0-3
Thread(s) per core:              2
Core(s) per socket:              2
Socket(s):                       1
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           79
Model name:                      Intel(R) Xeon(R) CPU @ 2.20GHz
Stepping:                        0
CPU MHz:                         2199.998
BogoMIPS:                        4399.99
Hypervisor vendor:               KVM
Virtualization type:             full
L1d cache:                       64 KiB
L1i cache:                       64 KiB
L2 cache:                        512 KiB
L3 cache:                        55 MiB
NUMA node0 CPU(s):               0-3```

Failure to load libcuda.so is normal, that's initializing the plugin registry

Hi there,

Are you using unique identities for each participant that’s connecting? In other words, when you generate a token for each participant, are you setting their identity field to something unique?

If this isn’t the issue, we’ll need some more information or a code snippet and/or log output to understand what’s going on. It’s difficult to help you debug from the description you shared.

congestion control disable solved the issue

Hm, on the surface that seems like an unintentional fix? cc <@U01KZ61GV5Z> 

As using iceTransportPolicy:true without no pass username solved it

Hey all. I’m running under docker locally. Is it possible to get livekit to log to a file rather than the default? I couldn’t see in the config-sample.yaml on how to do this. Thanks

Ok I just found

&gt;  we print logs into stderr/stdout, so it depends on where your docker output logs are being redirected do
So I will investigate that.

setup egress to debug mode and set ws url to wss:&lt;domain&gt;, keep the insecure to default

<@U045CJPACG1> I'm not using TLS. Why would I write wss?

ok then pass as <ws://domain>.
Egress service should be able to access into your livekit servers via outside (public internet), which will send a request  from <https://egress-composite.livekit.io> (default layout site). You should the egress service exactly error in egress service server logs
```"url": "<https://egress-composite.livekit.io?layout=speaker-dark&amp;token=>&lt;token&gt;&amp;url=wss%3A%2F%2F&lt;domain&gt;"}	```


My livekit servers are accessible from outside. My app using my livekit servers right now.

There is a config flag for `insecure` that needs to be set to true. Otherwise, chrome cannot connect to an insecure websocket

<https://github.com/livekit/egress/blob/main/pkg/config/base.go#L23>

it is already set. insecure: true . it's not working.

egress-helm.yaml

if your livekit-server is accessible from outside, it is using TLS. Have you tried using `wss://`?

rtcserver.yaml

I will try but this is my livekit helm yaml. I don't think I am using TLS.

```"error":"signal error dial error: tls: first record does not look like a TLS handshake"```
Because I am not using TLS.

The helm chart is meant to be used with TLS (see pre-requisites):
<https://docs.livekit.io/oss/deployment/kubernetes/#pre-requisites>

I’m not sure how you have `mydomain` mapped to a livekit instance otherwise, or whether there are firewall issues with connecting the two deployments

I set up a livekit server with the helm yaml file above and I can broadcast with the address <ws://mydomain:7880>.
I'm not using kubernetes on AWS, GCloud, or DO. How can I configure TLS?

pls help, i tried all the setup for windows for localhost setup but none of them working ?
can anyone pls guide me a lil which one to use and how

also deployed on ec2 and trying to connect from localhost react sdk but showing this:
error could not connect PeerConnection after timeout

```# turn server
# turn:
#   # Uses TLS. Requires cert and key pem files by either:
#   # - using turn.secretName if deploying with our helm chart, or
#   # - setting LIVEKIT_TURN_CERT and LIVEKIT_TURN_KEY env vars with file locations, or
#   # - using cert_file and key_file below
#   # defaults to false
#   enabled: false
#   # defaults to 3478 - recommended to 443 if not running HTTP3/QUIC server
#   # only 53/80/443 are allowed if less than 1024
#   udp_port: 3478
#   # defaults to 5349 - if not using a load balancer, this must be set to 443
#   tls_port: 5349
#   # set UDP port range for TURN relay to connect to LiveKit SFU, by default it uses a any available port
#   relay_range_start: 1024
#   relay_range_end: 30000
#   # set external_tl to true if using a L4 load balancer to terminate TLS. when enabled,
#   # LiveKit expects unencrypted traffic on tls_port, and still advertise tls_port as a TURN/TLS candidate.
#   external_tls: true
#   # needs to match tls cert domain
#   domain: <http://turn.myhost.com|turn.myhost.com>
#   # optional (set only if not using external TLS termination)
#   # cert_file: /path/to/cert.pem
#   # key_file: /path/to/key.pem```

no password and username configuration section, but client side requires username and password always

<https://docs.livekit.io/oss/deployment/#improving-connectivity-with-turn>

it's handled automatically with livekit clients

Hi Pat,
this should be possible by publishing additional tracks with `Source.Unknown` set on the publication/track!
You could use the `name` field in order to differentiate different tracks that you are publishing that way

Oh is that right! Time to investigate into this and see how welll I can implement this. Thanks <@U02EU2ZPTQU>

image.png

This warning results in other warning regarding using methods from LivekitReactNative class. like AudioSession

our sdk contains native modules, so expo go isn’t supported.

yea I realize that, I test with development builds using the expo simulator

are you building using eas build or locally?

I ended up installing livekit-client and going with @livekit/react-components.

<@U02S1919YLB> the reason for that is that in newer versions `livekit-client` is treated as a peer dependency (instead of a direct dependency of livekit-react).
So installing `livekit-client` was the right thing to do :slightly_smiling_face:

<@U02EU2ZPTQU> many thanks for the clarification!  :raised_hands:

`useRoom` itself doesn’t connect. the `useRoom`only returns the `connect` function among other things.

the `startAudio`  error is due to browser policies, there’s a static method on `Room` to manually call startAudio after a user gesture: <https://docs.livekit.io/client-sdk-js/classes/Room.html#startAudio>

Maybe the lock up you are seeing is connected to room being potentially undefined on first render in the latest version, can you double check that?
This was a recent change in order to fix auto-disconnects (related PR <https://github.com/livekit/livekit-react/pull/112>)

I’m guessing here, but I think room was undefined with 0.6.2 but with the latest it’s being defined but the state is disconnected and then room name is blank.  So maybe this is the difference?

I can change my code because a lot of it was based on room being undefined.

Screenshot 2023-01-04 at 13.14.41.png

So yes, to answer your question.  It’s undefined and then gets defined.  I didn’t check for this at the beginning.  I really should have.  It was only when I decided to comment out a lot of the code and leave the useRoom, that I found out the root cause.

Screenshot 2023-01-04 at 13.31.43.png

<@U02EU2ZPTQU> I decided to refactor code similar to the example provided.  Where all the logic is in it’s own component and initialised when I pass a token to it.

It now behaves a lot better than before.

One question I have now.  1.1.0 vs 1.0.3.

With 1.0.3 is no problem at all.  I can connect and disconnect and the microphone seems to work, I get the recording notification in the tab.  Seems fine.

1.1.0.  I get room is not ready yet and it does not connect.

Same code.

Did anything change significantly between 1.0.3 and 1.1.0?

Thanks

I can put my code in a gist if that’s helpful.

cc: <@U02EU2ZPTQU>.

<@U0444GJ7CLC> We do have tests that sets the quality explicitly and ensures that the desired quality is received. It should take effect within a few hundred ms. Are you seeing that it does not take effect at all? Which version of server are you using?

No, it doesnt take effect at all, I am using livekit cloud

can you share your livekit-client debug-level logs of the 1s pre and 2s post attempted layer switch?

I am using `@livekit/react-components`  there is no option to set loglevel in roomOptions. <@U02EU2ZPTQU>

you will have `livekit-client` installed as well (as a peer dependency) you can just `import {setLogLevel} from 'livekit-client'`

localhost-1672839743734.log

localhost-1672840198797.log

In the above log , there is one screenshare track also
In this log , there is no screenshare track , only camera track, for better debug <@U02EU2ZPTQU>

thanks! Unfortunately I don’t see anything out of the ordinary there right away.

You can try and reproduce it as well. There is definitely some issue here.

will try to reproduce, which browsers are you seeing this with (including version) for both publisher and subscriber?

*Publisher and Subscriber* - Google Chrome Version 108.0.5359.124 (Official Build) (arm64)

It was working as expected a month back, before I went on holidays :slightly_smiling_face:

Just tried to reproduce on cloud with the sample app in the `client-sdk-js` repo.
Steps taken
• disable `adaptiveStream`
• once two participants are connected, switch between different`PreferredQuality` settings
=&gt; Quality updates accordingly on the subscriber side.

my publisher only publishes 1 simulcast stream , original is `VideoPresets43.h480` and simulcast layer is `VideoPresets43.h240`

autoSubscribe is also disabled, I subscribe to tracks as and when required

Can this cause any problem? I subscribe and in the next line i set the video quality.

Yeah, I think you’re right, that it might be connected to `autoSubscribe` being off. We were able to reproduce it and updated cloud.
Could you try again now to see if it works?

Yes <@U02EU2ZPTQU> I can see the issue seems to be fixed now.

we'll double check your instance. there aren't any limitations as far as I know

I do not see any issues on our end regarding webhooks. did you change any configuration in terms of the keys used to authenticate the webhook event?

no and i recreated the webhook on Saturday with a different API key as well

receive a log on our end when calling the endpoint directly via `curl`

e.g.
```curl -i -X POST https://$HOST/webhooks/livestreaming-event```
produces the log

```Received a request to /webhooks/livestreaming-events
Warning: Received a request to /webhooks/livestreaming-events without an authorization header```
on our server

i am testing with this command which formerly produced events to a prior webhook configuration

```livekit-cli join-room \
--url <https://harbor.livekit.cloud|https://harbor.livekit.cloud> \
--room rooms/cameras/camera-0 \
--identity "cameras/camera-0" \
--publish-demo \
--api-key $LIVEKIT_API_KEY \
--api-secret $LIVEKIT_API_SECRET```

we've had issues with our TLS certificate before so im thinking that might be the issue. i was able to get my local webhook working again through HTTPS via Ngrok

to follow up here TLS was the issue. we are receiving LiveKit events to our webhook now!

The docker instance doesn’t work.  I get no feedback.

I just installed the brew version and running the same yaml file.

I only see

&gt; webhook {}
in the nest console.  I have outputted both @Body which is supposed to capture req.body and also getting the request and outputting req.body.

Same output.

```async webhook(@Req() req: RawBodyRequest&lt;Request&gt;) {
    const raw = req.rawBody; // returns a `Buffer`.
    console.log(raw);```
I now get undefined.

So I came across this:  <https://livekit-users.slack.com/archives/C01KVTJH6BX/p1641232292225400>

and verified nest has it:
<https://github.com/nestjs/nest/issues/10587#issuecomment-1324243918>

I have literally no idea what I am doing.  I don’t use nestjs at all.

Hopefully this helps someone.

My main.ts:

```import { NestFactory } from '@nestjs/core';
import { AppModule } from './app/app.module';
import { urlencoded, json } from 'express';
import * as bodyParser from 'body-parser';

async function bootstrap() {
  const app = await NestFactory.create(AppModule, {
    bodyParser: false,
  });

  const rawBodyBuffer = (req, res, buf: Buffer, encoding) =&gt; {
    console.log('buff', buf.toString());
    if (buf &amp;&amp; buf.length) {
      req.rawBody = buf.toString(encoding || 'utf8');
    }
  };

  app.use(bodyParser.json());
  app.use(bodyParser.text());
  app.use(bodyParser.urlencoded());

  app.use(
    bodyParser.raw({
      verify: rawBodyBuffer,
      type: ['*/*'],
    }),
  );

  app.enableCors();
  app.setGlobalPrefix('api');
  app.use(json({ limit: '50mb' }));
  app.use(urlencoded({ extended: true, limit: '50mb' }));
  await app.listen(process.env.LISTEN_PORT);
}
bootstrap();```

Hey <@U02S1919YLB>, just want to verify you got past this? Or still stuck here?

<@U01KZ0SCDKL> Thanks for responding!  I got past this.  The above should help any other devs who use nest.  As I see searching for similar posts.  Many other devs use express.

Appreciate you posting this here for other devs. Now I just gotta figure out how to get this content somewhere that’s searchable after Slack removes their Premium free trial. :thinking_face:

I'm currently using these videoGrants but users can still join the room even if it hasn't been created yet.

```"video": {
    "roomCreate": false,
    "roomJoin": true,
    "room": "my-test-room",
    "canSubscribe": true
  }```

My goal is making sure a room was created by a moderator / Account owner before allowing participants to join the room.

I think you can do this by disabling *<https://github.com/livekit/livekit/blob/master/config-sample.yaml#L130|auto_create>* in the config. Then rooms have to be explicitly created with `CreateRoom` before they become usable

Through the cloud, is this an option which can be applied per project?

Maybe this will require myself to build a validation script prior to joining, if this is not a thing in cloud.

Wow, I sure had missed that in the UI. My bad. Thanks as always <@U01KZ61GV5Z>!

This said, even if I uncheck this, a user can still join the room even if the room hasn't been created manually. <@U01KZ61GV5Z>, are you aware if there's a delay for this to work?

yes, there is a delay for this as the values are cached. I think it's cached for about 5-10 mins

That's somewhat what I was thinking. Wonderful. Ill give it another run in a bit. Thanks :slightly_smiling_face:

<@U01KZ61GV5Z>, could it be possible that the cache refreshing happens a lot less frequently?

I'm saying this because, I currently do not have any rooms part of my project, and the option `Rooms are automatically created when a participant joins` is unchecked in the Cloud UI, but participant joining with the videoGrants provided in the initial post can still join.

Hello <@U01KZ61GV5Z>, could it be possible that the cloud option has no effect? Or could this be a bug?

Just gave it another go with the following token this morning, and I was still able to join a non-created room.

```{
  "sub": "user01",
  "jti": "user01",
  "exp": 1673032347,
  "nbf": 1673017947,
  "iat": 1673017947,
  "iss": "APIHDb4NpGLEDNG",
  "name": "Test User",
  "video": {
    "roomCreate": false,
    "roomJoin": true,
    "room": "my-test-room",
    "canSubscribe": true
  }
}```


it's possible there's a bug with this setting. we'll investigate

Just coming back to this one, has any investigation started for this?

Hey <@U01KZ61GV5Z>, I was simply circling back to this issue, and I was wondering if you had any update?

Hey, sorry I do not. it's on the list of things we'll look into, but have a few other higher priority things to chase down at the moment

Hey Pat, this has been fixed and will be deployed to Cloud next week. Thank you for your patience here!

Capture.PNG

can you dump the ICE candidates table from <chrome://webrtc-internals> ?

webrtc_internals_dump.txt

webrtc_internals_dump (2).txt

webrtc_internals_dump (7).txt

webrtc_internals_dump (4).txt

webrtc_internals_dump (3).txt

webrtc_internals_dump (1).txt

webrtc_internals_dump (5).txt

webrtc_internals_dump (6).txt

Thank you so much David for taking the time to respond to my issue. I really appreciate your attention and support.

We are currently in the process of evaluating alternative solutions for our current service(wowza) and we came across LiveKit. It appears to be a good match for our needs and we would like to learn more about it.

pls let me know if I missed anything or anything else needed to know about the issue

Yes, in general, it should. Are you seeing otherwise?

hi <@U04CR47TT25>. what cloud provider are you using? there should be a log where you can see if there are errors in the script. see <https://stackoverflow.com/questions/18691867/where-to-find-logs-for-a-cloud-init-user-data-script>

Hi <@U02BBTRGF98> sorry for the late response didn't see the message! I am currently using digital ocean (created a droplet with ubuntu running on it) and I am now trying to run `init_script.sh` I am planning on doing the same on another cloud provider called akash but I decided to first try it on DigitalOcean, I used   `scp` to move my script to the droplet and I have write permissions however for some reason it still doesn't work could you help me with this?

No problem <@U04CR47TT25>! Digital Ocean with Ubuntu should work just fine. When you say it doesn’t work, could you provide some more info? Do you see any errors when running `init_scrip.sh` ? have you gone through the troubleshooting steps provided here? <https://docs.livekit.io/oss/deployment/vm/#troubleshooting>

and is there a way I could see if an error has occured?

can you confirm whether the video is being outputted to other participants in the room?

I noticed that my local video is only shown when I connect in some room. I wanted to show the preview before connecting

I think this might be possible if you create the video track manually and show the videoTrack directly on the video renderer before connecting

```val track = localParticipant.createVideoTrack()
track.startCapture()

// ...
// when connected
localParticipant.publishVideoTrack(track)```

Yes, I already do, I want to do the camera view without connecting the room

Even starting the capture is only rendered after connected

are you passing the track directly to the video renderer or are you trying to get it through the localParticipant’s track publications?

I get direct from room.localParticipant, and use the trackers.

``` val trackMap by participant::videoTracks.flow.collectAsState()

    val videoTrackMap = trackMap.minByOrNull {
        it.first.source == Track.Source.CAMERA
    }

    val videoTrack = videoTrackMap?.second as? VideoTrack

    if (videoTrack != null) {
        VideoRenderer(
            modifier = modifier,
            room = room,
            videoTrack = videoTrack,
        )
    }```

```    VideoItemTrackSelector(
        modifier = Modifier.fillMaxSize(),
        room = vm.room,
        participant = vm.localParticipant,
    )```

ah yeah, iirc, the video tracks will be empty until connected since it doesn’t have a valid publication yet

But, same when i published, only update when conect to room

```    suspend fun publishLocalCamera() {

        val videoTrack = localParticipant.createVideoTrack(
            options = LocalVideoTrackOptions(
                position = CameraPosition.FRONT,
                captureParams = VideoPreset43.QHD.capture,
            )
        )

        videoTrack.startCapture()

        localParticipant.apply {
            publishVideoTrack(videoTrack)

            setMicrophoneEnabled(true)
            setCameraEnabled(true)
        }
    }```

My expected behavior is to open the application and show it with the camera rendering, without being connected to any room

my suggestion is to cache the videoTrack yourself and pass it directly through to the videoRenderer, rather than looking for it through the localParticipant

Are you looking to automatically upload all tracks to Google Cloud? If so take a look at <https://docs.livekit.io/server/egress/autoegress/|Auto Egress docs>. This would work with Cloud as well.

Thank you. But how should I trigger the CreateRoom operation? I assume it's from the baceknd but I don't understand how to do it.

Hi <@U04HZHY03PW> you'll need to integrate LiveKit from your backend too. `CreateRoom` is a server API that's accessible from any one of our <https://github.com/livekit/livekit#server-sdks|server SDKs>

If I want the egress a specific track , what would be the trackID in the following js command:
egressClient.startTrackEgress('my-room', output, trackID);

How do I get this value?

You can get this in two ways:
• via client SDKs: when you publish a track, it'll return a <https://docs.livekit.io/concepts/api-primitives/#trackpublication-attributes|TrackPublication> object, which contains the `sid` of the track.
• via server SDKs: you can get a webhook callback whenever a track is published, or call `ListParticipant` API to dump all of that participant's tracks.

how to integrate livekit and camera filter
can suggest me some good camera filters (similar to bigo, or current live apps)

Hi <@U04HCRRP9C7> , sorry for the late reply.
Were you able to resolve this issue ?

<@U02MAJLGN65> No, i am unable to solve this issue. Can you please help me out with this?

Below errors logged in console, can you help now?

*<http://ATAudioSessionPropertyManager.mm:359|ATAudioSessionPropertyManager.mm:359>  FAILED to set property 1684431725 on AVAudioSession with error Error Domain=NSOSStatusErrorDomain Code=-50 "(null)"*

*AURemoteIO.cpp:1679 AUIOClient_StartIO failed (561015905)*

Hi <@U04HCRRP9C7> thanks for the info, these are great hints. I think it's related to CallKit, but I don't have a CallKit example handy to debug these issues. Any chance you could prepare a minimal reproducible example simulating your issue ? That would help speed up resolving this issue .

Otherwise I can prepare a CallKit example myself to debug these issues but I'm afraid it will take a few weeks since I have other things I'm working on.

This information might be useful <https://youtu.be/JB2MdcY1MKs?t=6m23s>

udp transport fail

The NAT traversal fails in my environment, can you help to check? The log is attahced.
The tcpdump packet capture shows there is no response from livekit server for the DTLS SSL hello.
The udp port 7882 port is listened in the livekit server.