Kubernetes deployment randomly losing networking

This is pretty bizarre, but for some reason, our ShinyProxy pods on K8s are just randomly unable to connect to anything over the network. It begins with an error about not being able to reach kubernetes.default, and upon exec-ing into the pod with kubectl, you can see that there is no network connection at all. You can’t ping any websites, you can’t update the apt repos. We would normally look to K8s to figure out what’s going on, but from the outside everything seems well. Additionally, ShinyProxy pods are the only pods in our cluster that are having this issue. I was wondering if anyone else is seeing this.

This is our dockerfile:

FROM openjdk:8-jre

ARG EURMBS_INSIGHT_REPO
ENV EURMBS_INSIGHT_REPO ${EURMBS_INSIGHT_REPO}

RUN mkdir -p /opt/shinyproxy/
RUN wget https://www.shinyproxy.io/downloads/shinyproxy-2.3.0.jar -O /opt/shinyproxy/shinyproxy.jar
COPY application.yml /opt/shinyproxy/application.yml
COPY start_shiny_proxy.sh /opt/shinyproxy/start_shiny_proxy.sh

RUN mkdir /etc/certs/
COPY cert.pem /etc/certs/cert.pem

WORKDIR /opt/shinyproxy/
ENTRYPOINT ["/bin/bash", "./start_shiny_proxy.sh"]

And start_shiny_proxy.sh looks like this:

#!/usr/bin/env bash
set -e

sed -i "s/{EURMBS_INSIGHT_SHINY_APP_VER}/${EURMBS_INSIGHT_SHINY_APP_VER-NA}/" application.yml
sed -i "s/{NACMBS_INSIGHT_SHINY_APP_VER}/${NACMBS_INSIGHT_SHINY_APP_VER-NA}/" application.yml

exec java -jar ./shinyproxy.jar
1 Like