-
Notifications
You must be signed in to change notification settings - Fork 111
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
Screen Images Freezes after 30ish minutes #39
Comments
Please, post your layout.conf.default file (redacted as needed).
Please, recreate the symptoms, then run the following in an SSH client
(e.g., putty), 'sudo /usr/bin/displaycameras status'. That yields
interesting information from the various omxplayer instances (via DBUS).
Thanks,
Andy
…On Sun, Feb 2, 2020 at 2:03 PM Nicarlo ***@***.***> wrote:
*Describe the bug*
After about 30-45 minutes of runtime the images freeze up.
I checked htop and there is no resource exhaustion happening however I do
see the omxplayer.bin rtsp running in the process list. If I check the
network traffic via something like iftop or vnstat you can see that the
traffic has dropped dramatically to the point where i feel comfortable
saying that the traffic received is related to the typicaly idle noise
you'd see.
Some things that I've tried to do to troubleshoot:
1. Attempted to restart the NVR to see if it was related to the NVR
2. Ran a single RTSP feed from the debug omxplayer mode as found in
the READM.md
3. Changed the power supply of the PI to ensure that it wasnt related
to power.
*To Reproduce*
Steps to reproduce the behavior:
1. Install per README.md
2. Configure my feeds in layout.conf.default
3. Run 'sudo systemctl restart displaycameras.service'
4. Wait about 30-45 minutes until the images freeze
5. Run 'sudo systemctl restart displaycameras.service' again to
restart the sessions.
*Expected behavior*
The images should not freeze up
*Raspberry Pi (please complete the following information):*
- Hardware model: Raspberry Pi 3 Model B
- OS: Rasbian Buster Lite
- Console: Text Console
- gpu memory split: 256MB
- displaycameras Version 0.8.3.3
**Display Configuration
- Display/Monitor resolution: 1440 x 900
- Window matrix: 4x4
- Feed resolution: 704 x 480, Compression H.264H, 15 FPS, 256/Kb/S Bit
Rate, Bit Rate Type CBR
*Camera information (please complete the following information):*
- Number of cameras: 4
- Camera models (if not Ubiquiti): Dalhua 2 x 4K cameras and 2 x 4MP
Cameras
- Managed
*Features enabled (please complete the following information):*
- Rotation: no
- Display Blanking: no
- Display Detection: no
*NVR information (please complete the following information):*
- NVR hardware Dahua N41B2P3 ( see
https://us.dahuasecurity.com/wp-content/uploads/2018/09/Datasheet_4MP_4K_8ch_4K_Network_Security_System_N488D63_v001_001_FINAL.pdf
)
- NVR UniFi Video server version: System Version: 3.216.0000000.0,
Build Date: 2018-06-05
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#39?email_source=notifications&email_token=AHIYIKL3Q7SFJDFR7PYORU3RA4KIXA5CNFSM4KO3NPZ2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IKNWUIA>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AHIYIKOZEHYW5K23TRZPJVLRA4KIXANCNFSM4KO3NPZQ>
.
|
Hey Andy, thanks for reaching out Here's the layout.conf.default
output of the sudo /usr/bin/displaycameras status
|
I'm having the same problem, it freezes every 2 hours. My setup is exacly the same as Nicarlo. Using a LOREX NVR with same type of rtsp stream. If I do a restart everything is good for another couple hours. Stream is at low res H.264 352x240 512kbps....i've tried multiple resolutions with the same result. |
That is the output of the sudo /usr/bin/displaycameras status command AFTER
your display has frozen? If so, it appears that omxplayer still thinks
it's playing your feeds. If you run that command multiple times in
relatively quick succession (after things freeze), do the stream time
positions increase?
The package really just wraps omxplayer in enough management to call is a
service. Your setup just runs four instances of omxplayer concurrently and
restarts them if they report as not playing. That's it.
The only "rule" you're breaking is your running downscaling. It doesn't
matter that 480 is just a little bigger than 450, it still forces real-time
downscaling in software. To test this hypothesis, you can hand code the
omxplayer option changes in the main script (to run cropped videos) or use
a 1080p display (and layout file).
Thanks,
Andy
…On Tue, Feb 4, 2020, 9:23 AM Nicarlo ***@***.***> wrote:
Hey Andy,
thanks for reaching out
Here's the layout.conf.default
# This is the camera feed/windows layout configuration file for the
# displaycameras service. It ONLY configures the layout and feeds for
# the cameras; the rest of the configuration is in displaycameras.conf.
# See the comments in that file for notes on configuring the below.
# This example defines seven 1/2-HD windows, three of which are off-screen to the right,
# through which the service rotates six camera feeds (it actually uses only six windows)
# on a full-HD monitor. If this suites your needs, modify only the camera names to taste
# and feed URLs to what your cameras or NVR provides.
# Window names
# 2x2 screen with 3 off-screen windows
windows=(upper_left upper_right lower_left lower_right)
# Make sure to account for each window above in the list below.
# Windows positions
window_positions=(
#First Row
#upper_left
# 960x540
"0 0 719 449" \
#upper_right
# 960x540
"720 0 1439 449" \
#Second Row (missing all but the far right window because large_left is double size
#lower_left
# 960x540
"0 450 719 899" \
#lower_right
# 960x540
"720 450 1439 899" \
#off-screen
# 960x540 window just off-screen to the right
"1920 0 2879 539" \
# 960x540 window just below the other
"1920 540 2879 1079" \
# 960x540 window just off-screen to the left
"2880 0 3839 539" \
)
# Camera Names
camera_names=(NE SE South SW )
# Make sure to account for each camera above in the list of feeds below.
# Camera Feeds
camera_feeds=( \
# Mid-Res if your RPi can handle the load
# "rtsp://xxx.xxx.xxx.xxx/yyyyy_1" \
# Low-Res otherwise
# "rtsp://xxx.xxx.xxx.xxx/yyyyy_2" \
#Front Door
***@***.***:554/cam/realmonitor?channel=1&subtype=1" \
#Driveway
***@***.***:554/cam/realmonitor?channel=2&subtype=1" \
#Side Door
***@***.***:554/cam/realmonitor?channel=3&subtype=1" \
#Street View
***@***.***:554/cam/realmonitor?channel=6&subtype=1" \
#West
"<Camera stream URL>" \
#Vestibule
"<Camera stream URL>" \
)
# Are we rotating cameras through the window matrix? (default false if not set here)
# rotate="true"
output of the sudo /usr/bin/displaycameras status
***@***.***:~ $ sudo /usr/bin/displaycameras status
NE is Playing ***@***.***:554/cam/realmonitor?channel=1&subtype=1
at 15485sec
SE is Playing ***@***.***:554/cam/realmonitor?channel=2&subtype=1
at 15726sec
South is Playing ***@***.***:554/cam/realmonitor?channel=3&subtype=1
at 15720sec
SW is Playing ***@***.***:554/cam/realmonitor?channel=6&subtype=1
at 15713sec
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#39?email_source=notifications&email_token=AHIYIKKL7SH2J537LFBQ6EDRBF25TA5CNFSM4KO3NPZ2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKXZN3Y#issuecomment-581932783>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AHIYIKPV2LVR4ALTIGI3GOLRBF25TANCNFSM4KO3NPZQ>
.
|
the After reading your post I started instead focusing my attention on omxplayer. After making changes to the resolution of the substream on the NVR, the feeds started playing again without me having to restart the session. @chienlaid try playing around with the resolution on your lorex NVR and let us know if this also fixes your issue |
A little update on this. I must have spoken to soon because the issue started happening again but this time after a longer time. I did ran the following command directly to get a log output of the omxplayer instance for a single feed. As @Anonymousdog mentioned his script is only a wrapper to make it easier to manage the omxplayer so I am convinced this is an issue with the omxplayer. For those interested here is the last few lines of the log output before the feed freezes
Additionally I wanted to rule out that it wasnt an issue with my NVR so I initiated a stream from another computer using VLC and left it running overnight. When I returned the stream was still working properly which furthers my thinking that its an issue with omxplayer. Lastly, I had a raspberry pi 4 4GB version laying around so I loaded that unit with raspbian lite and went through the installation of the displaycamera script and copied the old layout.default configuration to this unit. After some time, I still continue get a frozen feed. some poking around on the omxplayer github issue page I've come across quite a few threads with people reporting similar issues. I've listed some of them below but have not yet tried them all https://stackoverflow.com/questions/37053495/omxplayer-freezes-when-playing-video I will continue troubleshooting this issue and will let you know what I do find. |
Alright another update on this. The current maintainer of omxplayer is sayingthat omxplayer will be depreciated over vlc Now I did try running the same rtsp stream over vlc and I had no issues with it. |
Not saying I won't do it, but I have no plans.
…On Fri, Feb 7, 2020 at 8:42 PM Nicarlo ***@***.***> wrote:
Alright another update on this.
The current maintainer of omxplayer is sayingthat omxplayer will be
depreciated over vlc
see popcornmix/omxplayer#752 (comment)
<popcornmix/omxplayer#752 (comment)>
Now I did try running the same rtsp stream over vlc and I had no issues
with it.
@Anonymousdog <https://github.com/Anonymousdog> , any plans to do the
same as you did for omxplayer with vlc?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#39?email_source=notifications&email_token=AHIYIKJFGOFZBZOA6JNJJ33RBYEZPA5CNFSM4KO3NPZ2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELFFSJA#issuecomment-583686436>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AHIYIKP2ZWTGB4I2EPLGJH3RBYEZPANCNFSM4KO3NPZQ>
.
|
@Nicarlo - Have you had any luck since tinkering with this back in Feb? I am experiencing the same issue you described above. The image freezes and it takes a restart of the service to get it back online.
I have yet to try tinkering with any of the timing settings, could those make a difference? I currently have a cron job that just restarts the service every 30 minutes. But I still seem to catch the stream frozen from time to time. @Anonymousdog - Thank you for taking the time to put this together. It is a very neat solution. I would be very interested if you do the same with VLC. |
@AnonWSU86 I ended up settling for a single stream on vlc. Vlc does have the ability to do something similar however their documentation has gaps in it and no one appears to know how to get it working. As @Anonymousdog mentioned this isn’t an issue with displaycameras but instead an issue with the omxplayer that it uses. I did check out their gitlab and omxplayer will eventually get depreciated for vlc so there are no plans for them to put out a fix for it. |
Describe the bug
After about 30-45 minutes of runtime the images freeze up.
I checked htop and there is no resource exhaustion happening however I do see the omxplayer.bin rtsp running in the process list. If I check the network traffic via something like iftop or vnstat you can see that the traffic has dropped dramatically to the point where i feel comfortable saying that the traffic received is related to the typicaly idle noise you'd see.
Some things that I've tried to do to troubleshoot:
To Reproduce
Steps to reproduce the behavior:
Expected behavior
The images should not freeze up
Raspberry Pi (please complete the following information):
**Display Configuration
Camera information (please complete the following information):
Features enabled (please complete the following information):
NVR information (please complete the following information):
The text was updated successfully, but these errors were encountered: