Here is my Dockerfile, I add hadoop into the image:
FROM tensorflow/serving:2.10.0-gpu
RUN apt update && apt install -y openjdk-8-jre && apt-get clean
COPY hadoop-2.10.2 /root/hadoop
ENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64
ENV HADOOP_HDFS_HOME /root/hadoop
ENV LD_LIBRARY_PATH ${LD_LIBRARY_PATH}:${JAVA_HOME}/jre/lib/amd64/server
EXPOSE 8500
EXPOSE 8501
RUN echo '#!/bin/bash \n\n\
tensorflow_model_server --port=8500 --rest_api_port=8501 \
--model_name=${MODEL_NAME} --model_base_path=${MODEL_BASE_PATH}/${MODEL_NAME} \
"$@"' > /usr/bin/tf_serving_entrypoint.sh \
&& chmod +x /usr/bin/tf_serving_entrypoint.sh
ENTRYPOINT ["/usr/bin/tf_serving_entrypoint.sh"]
It can work with image tensorflow/serving:2.10.0, but raise exception with gpu image tensorflow/serving:2.10.0-gpu.
Here is the official Dockerfile-gpu: https://github.com/tensorflow/serving/blob/master/tensorflow_serving/tools/docker/Dockerfile.gpu
I can't figure out the issue why gpu image does't work.
I have tried the newer image versions, but all of them don't work.
It is sure that this is something about the image...
