-
-
Notifications
You must be signed in to change notification settings - Fork 30k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Errors with stream component and HA hangs up #22840
Comments
I saw a freeze myself yesterday however I did not see any errors at all. Glad you got an error as I was trying to reproduce it. When my freeze happened it was under different circumstances. I had one of my amcrest cameras with preload stream and one of my blue iris generic cameras as well on preload. I noticed after about 12 hours that the amcrest camera was stuck on the same timestamp, the video was no longer updating no more how long I kept it open or attempted to view the stream in. I then unchecked preload stream on both cameras. After about 5 minutes when the keep alive time should've stopped. My HA froze and I could not control any lights. Systemctl also showed HA up and running but there were no errors in the logs at the time. If I can reproduce it and I get logs I will post them here myself. I am also on 0.91.1 with HA running in a venv on an i5. CC: @hunterjm FYI |
My camera stream also stuck on the same timestamp few minutes before HA hangs. |
Same here but only with dafang camera the other camera works fine. Noticed that the problem came out when HA try to stop rtsp server. |
Same issue here with both dafang camera and mijia 1080p camera. Also my first hanging homeassistant since two years. |
I had the same issues with the frontend locking up. Didn't resolve until I turned off stream and uninstalled libavformat-dev and libavdevice-dev. Have been running stable since. The logs during the lockup didn't seem to show any issues, the GUI was just unresponsive. |
Do any of you happen to be running any other components that depend on FFMPEG? Even if they don’t support stream. I.e. Ring |
My setup is just 4 Foscam cameras of varying ages and an older Foscam camera I had to set up as mjpeg |
I have 4 cameras working on ffmpeg. |
I have Ring setup |
I have only one camera and no other component using ffmpeg. |
Is the issue consistent (happens frequently)? Is there any indication as to what you do before the error appears? |
In my experience (and I only kept it running for a day or so), the issue happened multiple times during the day. I couldn't find any rhyme or reason as to why. In some cases, I was actually using it and it just stopped responding and in others, I'd go back to it after some time and realize I just couldn't get in. |
Personally I have only seen this happen once and have yet to reproduce it. I started keeping a preloaded stream again on the camera that was stuck on the timestamp so time will tell if that issue comes back for me. |
I would add that if there's a certain level of logging I can turn on to try to gather data for this, I'm happy to restore the configuration that was failing and collect whatever logs are necessary when there's a lockup. |
I have had this same thing happen on several occasions and I can probably reproduce it -- I'm trying to preload a generic camera stream from a balky Hikvision wifi doorbell that occasionally will start lagging and go offline/restart. When that happens it now seems bring down Home Assistant completely. On each occasion absolutely nothing pertinent has been in the log file. I run default logging level at info, but I can try turning it up. Is there any other components I need to include besides these?
|
It sounds like the general theme is cameras with flaky connections that we are currently processing. Likely due to not properly handling dropped connections. That gives me something to work off of and reproduce, but all my cameras are hardwired, and I haven't had any issues... so I will need to force a disconnect 😆 |
@hunterjm you are spot on with that assessment with the stream dropping. I did a very simple test just now. I loaded up a stream in my dev HA instance and after 1 minute I pulled the power on the camera. I waited 30 seconds and plugged it back in. HA was frozen from that point on and nothing would load, not even the system logs. Here are the logs from this experience. https://hastebin.com/amuwedazeq.cs This is what HA looks like when it froze and I tried to look at the logs in the UI |
Same issue here. with preload enabled I get this https://youtu.be/i3HQevp-M3A and eventually HA hangs hard. (#22941 looks like a likely dupe of this issue) |
In my experience it's enough to have the camera entity visible for a while for HA to hang. It's not isolated to the more-info popup. A bit anecdotal, but could be relevant: |
I think we have fix #22982 |
I don't think this is fixed. I"m on 0.92.0 and experiencing the same thing if I have a preloaded stream. |
I still have the same problem on 0.92,0 |
The problem still persists on 0.92. Today one of my camera's got disconnected (the power cable came lose) and hass became unresponsive afterwards. All I had in logs after that was:
For each and every component on hass. Back to 0.89.2 until this is resolved. |
With my RTSP stream over WIFI I do not experience any complete HA stalls. But the Lovelace stream view of a Generic IP Camera entity stalls regularly and after that does not recover until I restart HA. My system is Hassio 0.92.2 on a Raspi3B. Interesting detail: VLC on Mac and on Android does not have any problems with stream outages in the RTSP stream. It just shows some block artifacts and continues to show the video. But tinyCam on Android in fact stalls in a very similar way as the HA Generic IP Cam does. Both VLC and tinyCam are built on FFMPEG. One difference I can see though, is that VLC on Android never tries to write the stream to an MP4 file, which tinyCam can do on request. So maybe there is a difference in the way FFMPEG is integrated that causes these problems when streams are WIFIed and thus unstable. |
Hassio hangs seems like it is due to Streaming an RTSP stream (foscam) resulting in Besides this I have endless ConnectionResetErrors ('Cannot write to closing transport') but I have the impression this is not causing too much trouble.
|
I believe this is caused by cameras losing connectivity, which might be solved by #23617. If I can have help testing after beta is cut, that would be great. |
Yes sure! |
@b0c1 - As I stated before, feel free to debug your situation locally and open a PR with a fix. I would be happy to review it. Some issues you might run in to:
|
@hunterjm just an idea but I think you can use pts instead of dts like the pyav frame seek example |
They are both |
The issue remains on 0.101.3:
|
Still persists in 0.102.2
|
still persistant in 0.104.1 |
I've the same problem with a Dlink dcs5222. My rtsp command line is running fine in VLC but no success in HA latest version. Very frustrating and I've never found a workaround for weeks. I just success to view the live video through mjpeg but loosing high quality. |
also have issue, stream via iSPY. HA version 0.104.3. Feb 04 00:41:40 homeassistant hass[19711]: 2020-02-04 00:41:40 INFO (MainThread) [homeassistant.components.stream] Started stream: http://192.168.13.10:8080/mjpegfeed?oid=1&full |
Having this issue as well (using the danfang hacks firmware):
Config: - platform: ffmpeg
name: "Xiaofang 1S"
input: !secret xiaofang1s_camera_rtsp_stream It was previously (and is still) working fine, meaning the video is correctly shown but the errors in the log are annoying... v0.105.5 (latest as of now) Any chance that this will be fixed? |
I get te following error with trying to show the amcrest doorbell camera (which is working in VLC player.
|
Know that the password what was in your post (that you have edited) also ends up in an notification e-mail: |
Hmmm... ok... I still don't have it working. Really weird |
well... it works. |
this still exists on 0.108.1. I am on pi 4 and noticed that homeassistant doesn't go up even after >5 mins. I noticed that container is up and resources are normal. Couple of container restart doesnt resolve it. After commenting stream:, hass is up. The strange part is, I was able to restart it successfully a couple of times before experimenting with camera.record service. |
Similar related issue, in LoveLace UI any live camera feed I show stops anywhere from 10 seconds > 30 minutes in, with |
There hasn't been any activity on this issue recently. Due to the high number of incoming GitHub notifications, we have to clean some of the old issues, as many of them have already been resolved with the latest updates. |
Same issue with a Trendnet camera (no DTS in packet error message), RTSP stream won't display or camera.record. It used to display in Lovelace before recent updates, and still displays just fine in VLC. |
I'm on HA version 0.116.2 and this is still a thing.. I'm using 3 Wyze cams (2x V2, 1x Pan) with the RTSP firmware and I still see these errors over and over in my logs: Timestamp overflow detected: dts = 117126, resetting stream Error demuxing stream: No dts in consecutive packets Stream packet without dts detected, skipping... Any chance this will ever get fixed ? |
Seeing the same here with 6x reolink cameras. All on gigabit wired links so it's not a poor connectivity issue. The stream component is all but useless at this stage, never a reliable camera feed, constant hangups, loading wheel, streams stuck in a loop since yesterday etc. Very unreliable in general. |
please reopen... |
Instead of re-opening this ticket (which at this point is over a year old and contains MANY different error messages), let's create specific individual issues for each distinctive error message in the log file. |
The errors usually show in the logs together so presumably they’re all related. Individual issues per specific error might only give half the picture. |
As the original developer and one of the code owners for the stream integration, I assure you that separate issues unless logs are identical would be much more beneficial than reviewing a year old issue. Especially since there have been a lot of modifications to the integration since. |
The error message is correct. There is no DTS in the packet. VLC and others don't mind but the component in home assistant does. |
Hello to all I'm also facing similar issue with the following log Logger: libav.rtsp RTP: PT=60: bad cseq 5bdd expected=4f2a And As you mentioned I opened a Issue but the issue is now closed without any way to solve this .... At last I re install my HA based on the Hassio OS but the problem keeps Can anyone please help me in order to solve this Thank you for your support and best regards |
Hello Can you please help me in this issue? If needed I will open a new issue again. Thank you and best regards |
Home Assistant release with the issue:
0.91.1
Last working Home Assistant release (if known):
None
Operating environment (Hass.io/Docker/Windows/etc.):
RPi 3B/Raspbian/Python 3.7.2/venv
Component/platform:
stream, camera/generic
Description of problem:
I use generic camera and stream component. When I tap on camera card in Lovelace on iPad (iOS app 1.5) I have errors in log, stream don't start and sometimes after that my HA hangs up. The frontend doesn't work, the light control automation doesn't start, systemctl shows that the HA service is active (running). I have to restart HA to using it. This is the first time HA has been hanging since I use it (2 years).
Problem-relevant
configuration.yaml
entries and (fill out even if it seems unimportant):Traceback (if applicable):
The text was updated successfully, but these errors were encountered: