test.yaml: update ci

- use ubuntu 20.04
- set nf_conntrack_max to avoid CrashLoopBackOff for kube proxy (see
kubernetes-sigs/kind#2240 (comment))
- print cluster info as soon as KinD is up

Signed-off-by: Mattia Mazzucato <matt.mazzucato@gmail.com>

.github/workflows/test.yaml

@@ -23,7 +23,7 @@ on:
 
 jobs:
   test:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       matrix:
         testSuite:
@@ -47,6 +47,11 @@ jobs:
         kustomize: 3.5.5
         helm: 2.16.7
         helmv3: 3.2.1
+    - name: Set nf_conntrack_max value
+      # This step is required to avoid CrashLoopBackOff for kube-proxy
+      # see https://github.com/kubernetes-sigs/kind/issues/2240#issuecomment-838510890
+      run: |
+        sudo sysctl net/netfilter/nf_conntrack_max=131072
     - uses: container-tools/kind-action@v1
       with:
         version: "v0.8.1"
@@ -55,6 +60,9 @@ jobs:
       run: |
         kubectl cluster-info
         kubectl get pods -n kube-system
+    - name: Dump cluster info
+      run: |
+        kubectl cluster-info dump
     - name: Setup needed resources
       run: |
         kubectl create namespace astarte-test