deploy: wait until pods are up and snapshotter added the CRD

This is crucial because otherwise creating the snapshotter class can
fail (depending on timing).

This is also useful because the caller then can be sure that the
driver installation has succeeded, without having to know details
about the pods that belong to the driver.

Author: pohly

README.md

@@ -53,6 +53,9 @@ statefulset.apps/csi-hostpath-provisioner created
         using           image: quay.io/k8scsi/csi-snapshotter:v1.0.1
 service/csi-hostpath-snapshotter created
 statefulset.apps/csi-hostpath-snapshotter created
+23:04:27 waiting for hostpath deployment to complete, attempt #0
+23:04:38 waiting for hostpath deployment to complete, attempt #1
+23:04:48 waiting for hostpath deployment to complete, attempt #2
 deploying snapshotclass
 volumesnapshotclass.snapshot.storage.k8s.io/csi-hostpath-snapclass created
 ```

deploy/deploy-hostpath.sh

@@ -138,6 +138,24 @@ for i in $(ls ${BASE_DIR}/hostpath/*.yaml | sort); do
     done | kubectl apply -f -
 done
 
+# Wait until all pods are running. We have to make some assumptions
+# about the deployment here, otherwise we wouldn't know what to wait
+# for: the expectation is that we run attacher, provisioner,
+# snapshotter and hostpath plugin in the default namespace.
+cnt=0
+while [ $(kubectl get pods 2>/dev/null | grep '^csi-hostpath.* Running ' | wc -l) -lt 4 ] || ! kubectl describe volumesnapshotclasses.snapshot.storage.k8s.io 2>/dev/null >/dev/null; do
+    if [ $cnt -gt 30 ]; then
+        echo "Running pods:"
+        kubectl describe pods
+
+        echo >&2 "ERROR: hostpath deployment not ready after over 5min"
+        exit 1
+    fi
+    echo $(date +%H:%M:%S) "waiting for hostpath deployment to complete, attempt #$cnt"
+    cnt=$(($cnt + 1))
+    sleep 10
+done
+
 # deploy snapshotclass
 echo "deploying snapshotclass"
 kubectl create -f ${BASE_DIR}/snapshotter/csi-hostpath-snapshotclass.yaml