Load balancer example check improvement
The load balancer example check script is not verifying the traffic is passing and distributed by the balancer. We need a more sophisticated check that will ensure the balancer is operating as expected.
@uablrek can you please check if we can have a more sophisticated check?
Yes I will try. Thanks for fixing the ci problems, I can't say I have understood those. Now when it works I can try to extend the testing. My problem has been that the base-image for NSE lacks every tcp test tool I could think of, only "ping" exists, so test from within the NSE pod is hard. I will try to test from main netns on nodes with "nc".
Oh, you can install whatever you need even in the check script if you want to.
If it is Alpine based it is a matter of apk update and then apk install curl for example. Ubuntu is apt-get update && apt-get install -y curl and so on. What tools do you need?
Tried to add install of nc in the Dockerfile;
FROM ${VPP_AGENT} as runtime
RUN apt install -y netcat-openbsd
...
But unfortunately any "apt install" fails in the "${VPP_AGENT}" image;
# apt-get install -y netcat-openbsd
Reading package lists... Done
Building dependency tree
Reading state information... Done
E: Unable to locate package netcat-openbsd
I tried a number of packages and all fails.
I will try something else.
So the images get distributed with cleaned package cache to make time lighter weight. The typical way to make this work is like this: RUN apt-get update && apt-get install -y netcat-openbsd And then it is a good practice to clean up the cache again before the final image is built. So the final execution might be something like:
RUN apt update && apt install -y netcat-openbsd && rm -rf /var/lib/apt/lists/*
Seems to work fine. Thanks!