Headless mode fails with EGL error

Accelerated Computing Intelligent Video Analytics DeepStream SDK
,
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
Jetson AGX Orin 64G

• DeepStream Version
7.0

• JetPack Version (valid for Jetson only)

R36 (release), REVISION: 3.0, GCID: 36923193, BOARD: generic, EABI: aarch64, DATE: Fri Jul 19 23:24:25 UTC 2024

• TensorRT Version
6.0+b106 / 8.6.2.3-1+cuda12.2

• NVIDIA GPU Driver Version (valid for GPU only)
N/A

• Issue Type( questions, new requirements, bugs)
Question

• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

Issue: when running GST_DEBUG=3 deepstream-app -c setup2.txt we’re getting EGL display error:

nvbufsurftransform: Could not get EGL display connection
...
nvbufsurface: Could not get EGL display connection
nvbufsurface: Can't get EGL display
0:00:06.118061411  1453 0xaaaac4f57470 ERROR       nvinferallocator gstnvinfer_allocator.cpp:107:gst_nvinfer_allocator_alloc: Error: Could not map EglImage from NvBufSurface for nvinfer
0:00:06.118086019  1453 0xaaaac4f57470 WARN              GST_BUFFER gstbuffer.c:958:gst_buffer_new_allocate: failed to allocate 88 bytes
0:00:06.118104035  1453 0xaaaac4f57470 WARN              bufferpool gstbufferpool.c:318:do_alloc_buffer:<bufferpool0> alloc function failed
0:00:06.118114147  1453 0xaaaac4f57470 WARN              bufferpool gstbufferpool.c:351:default_start:<bufferpool0> failed to allocate buffer
0:00:06.118121955  1453 0xaaaac4f57470 ERROR             bufferpool gstbufferpool.c:572:gst_buffer_pool_set_active:<bufferpool0> start failed
0:00:06.118133699  1453 0xaaaac4f57470 WARN                 nvinfer gstnvinfer.cpp:1008:gst_nvinfer_start:<primary_gie> error: Failed to set buffer pool to active
0:00:06.143608591  1453 0xaaaac4f57470 WARN                GST_PADS gstpad.c:1141:gst_pad_set_active:<primary_gie:sink> Failed to activate pad
** ERROR: <main:706>: Failed to set pipeline to PAUSED
Quitting
ERROR from primary_gie: Failed to set buffer pool to active
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(1008): gst_nvinfer_start (): /GstPipeline:pipeline/GstBin:primary_gie_bin/GstNvInfer:primary_gie
App run failed

But:

  1. This’s going to run in a docker w/o display.
  2. DISPLAY is unset - confirmed:
root@b06176e75a5a:/test# env | grep DISPLAY
root@b06176e75a5a:/test#
  1. the config file (setup2.txt) is minimal and does not even involve OSD/EGLSink:
[application]
enable-perf-measurement=0

[source0]
enable=1
#Type - 1=CameraV4L2 2=URI 3=MultiURI 4=RTSP
type=3
uri=file://./test.mp4
num-sources=1
gpu-id=0
cudadec-memtype=0


[sink0]
enable=1
gpu-id=0
type=3
#1=mp4 2=mkv
container=1
#1=h264 2=h265
codec=1
#encoder type 0=Hardware 1=Software
enc-type=0
sync=0
#iframeinterval=10
bitrate=2000000
#H264 Profile - 0=Baseline 2=Main 4=High
#H265 Profile - 0=Main 1=Main10
# set profile only for hw encoder, sw encoder selects profile based on sw-preset
profile=0
output-file=out.mp4
source-id=0

[streammux]
gpu-id=0
##Boolean property to inform muxer that sources are live
live-source=0
batch-size=1
##time out in usec, to wait after the first buffer is available
##to push the batch even if the complete batch is not formed
batched-push-timeout=40000
## Set muxer output width and height
width=1920
height=1080
##Enable to maintain aspect ratio wrt source, and allow black borders, works
##along with width, height properties
enable-padding=0
nvbuf-memory-type=0
## If set to TRUE, system timestamp will be attached as ntp timestamp
## If set to FALSE, ntp timestamp from rtspsrc, if available, will be attached
# attach-sys-ts-as-ntp=1




# config-file property is mandatory for any gie section.
# Other properties are optional and if set will override the properties set in
# the infer config file.
[primary-gie]
enable=1
gpu-id=0
model-engine-file=model_b1_gpu0_fp32.engine
#Required to display the PGIE labels, should be added even when using config-file
#property
batch-size=1
#Required by the app for OSD, not a plugin property
bbox-border-color0=1;0;0;1
bbox-border-color1=0;1;1;1
bbox-border-color2=0;0;1;1
bbox-border-color3=0;1;0;1
interval=0
#Required by the app for SGIE, when used along with config-file property
gie-unique-id=1
nvbuf-memory-type=0
config-file=bin2.txt


[tests]
file-loop=0

Any help?

2

some more info. I’ve tried to use the sample config:
source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

after modifying its sink to use FakeSink instead of EGL (ie. type=2 → type=1) I’ve tried to run it: /opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app# GST_DEBUG=3 deepstream-app -c source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

which also fails w same error:

nvbufsurftransform: Could not get EGL display connection
Setting min object dimensions as 16x16 instead of 1x1 to support VIC compute mode.
0:00:05.577941647  1584 0xaaaacf049490 INFO                 nvinfer gstnvinfer.cpp:682:gst_nvinfer_logger:<secondary_gie_1> NvDsInferContext[UID 5]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:2095> [UID = 5]: deserialized trt engine from :/opt/nvidia/deepstream/deepstream-7.0/samples/configs/deepstream-app/../../models/Secondary_VehicleMake/resnet18_vehiclemakenet.etlt_b16_gpu0_int8.engine
INFO: [Implicit Engine Info]: layers num: 2
0   INPUT  kFLOAT input_1         3x224x224
1   OUTPUT kFLOAT predictions/Softmax 20x1x1

0:00:05.890315777  1584 0xaaaacf049490 INFO                 nvinfer gstnvinfer.cpp:682:gst_nvinfer_logger:<secondary_gie_1> NvDsInferContext[UID 5]: Info from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2198> [UID = 5]: Use deserialized engine model: /opt/nvidia/deepstream/deepstream-7.0/samples/configs/deepstream-app/../../models/Secondary_VehicleMake/resnet18_vehiclemakenet.etlt_b16_gpu0_int8.engine
nvbufsurface: Could not get EGL display connection
nvbufsurface: Can't get EGL display
0:00:05.900967228  1584 0xaaaacf049490 ERROR       nvinferallocator gstnvinfer_allocator.cpp:107:gst_nvinfer_allocator_alloc: Error: Could not map EglImage from NvBufSurface for nvinfer
0:00:05.900991452  1584 0xaaaacf049490 WARN              GST_BUFFER gstbuffer.c:958:gst_buffer_new_allocate: failed to allocate 88 bytes
0:00:05.901009244  1584 0xaaaacf049490 WARN              bufferpool gstbufferpool.c:318:do_alloc_buffer:<bufferpool0> alloc function failed
0:00:05.901017116  1584 0xaaaacf049490 WARN              bufferpool gstbufferpool.c:351:default_start:<bufferpool0> failed to allocate buffer
0:00:05.901023164  1584 0xaaaacf049490 ERROR             bufferpool gstbufferpool.c:572:gst_buffer_pool_set_active:<bufferpool0> start failed
0:00:05.901033916  1584 0xaaaacf049490 WARN                 nvinfer gstnvinfer.cpp:1008:gst_nvinfer_start:<secondary_gie_1> error: Failed to set buffer pool to active
0:00:05.906676283  1584 0xaaaacf049490 WARN                GST_PADS gstpad.c:1141:gst_pad_set_active:<secondary_gie_1:sink> Failed to activate pad
** ERROR: <main:706>: Failed to set pipeline to PAUSED
Quitting
nvstreammux: Successfully handled EOS for source_id=0
nvstreammux: Successfully handled EOS for source_id=1
nvstreammux: Successfully handled EOS for source_id=2
nvstreammux: Successfully handled EOS for source_id=3
ERROR from secondary_gie_1: Failed to set buffer pool to active
Debug info: /dvs/git/dirty/git-master_linux/deepstream/sdk/src/gst-plugins/gst-nvinfer/gstnvinfer.cpp(1008): gst_nvinfer_start (): /GstPipeline:pipeline/GstBin:secondary_gie_bin/GstNvInfer:secondary_gie_1
App run failed
4

Have you run the unset DISPLAY command?

5

yes - as mentioned above, since this’s the suggestion so common :)

    DISPLAY is unset - confirmed:

root@b06176e75a5a:/test# env | grep DISPLAY
root@b06176e75a5a:/test#
6

The default DISPLAY maybe not in the env. Could you run the cat /proc/*/environ 2>/dev/null | tr '\0' '\n' | grep '^DISPLAY=:' | uniq command to check?

You can also change the sink to fakesink and have a try.

7

Please read my post again:

after modifying its sink to use FakeSink instead of EGL (ie. type=2 → type=1)

No offense, but as this’s the second time, please please read my entire post carefully for what I had tried already.

And as mentioned, I didn’t just run the env - I literally ran unset DISPLAY, checked docker run in process, checked proc, checked X11 fwding - nope, negative, nil, nothing.

root@b06176e75a5a:/# cat /proc/*/environ | grep DISPLAY
root@b06176e75a5a:/# 
root@b06176e75a5a:/# cat /proc/*/environ 2>/dev/null | tr '\0' '\n' | grep '^DISPLAY=:' | uniq
root@b06176e75a5a:/# 
root@b06176e75a5a:/#
8

Sorry. At present, we only support such scenario on the dgpu without a monitor, and this is not supported on jetson.

9

That…somehow doesnt sound right.

When you say “such scenarios”, are you suggesting that Jetson units can’t run deepstream pipelines (w nvinfer as one of element) w/o a monitor?

11

Hi,
If your device is headless, please make sure to apply this:

Jetson AGX Orin FAQ
Q: How to run developer kit in headless mode?

So that display is completely disabled.

12

Hi @fang4 , is the problem fixed now?

13

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

14

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.