From 758e650a280803bbe56469b78123cc37250e33c3 Mon Sep 17 00:00:00 2001
From: Hyounggyu Choi <Hyounggyu.Choi@ibm.com>
Date: Thu, 15 Aug 2024 13:37:11 +0200
Subject: [PATCH] tests: Ensure exec_host() consistently captures command
 output

The `exec_host()` function often fails to capture the output of a given command
because the node debugger pod is prematurely terminated. To address this issue,
the function has been refactored to ensure consistent output capture by adjusting
the `kubectl debug` process as follows:

- Keep the node debugger pod running
- Wait until the pod is fully ready
- Execute the command using `kubectl exec`
- Capture the output and terminate the pod

This commit refactors `exec_host()` to implement the above steps, improving its reliability.

Fixes: #10081

Signed-off-by: Hyounggyu Choi <Hyounggyu.Choi@ibm.com>
---
 tests/integration/kubernetes/tests_common.sh | 52 +++++++++++---------
 1 file changed, 29 insertions(+), 23 deletions(-)

diff --git a/tests/integration/kubernetes/tests_common.sh b/tests/integration/kubernetes/tests_common.sh
index a93022d8aa..c88b4adecd 100644
--- a/tests/integration/kubernetes/tests_common.sh
+++ b/tests/integration/kubernetes/tests_common.sh
@@ -76,17 +76,17 @@ get_one_kata_node() {
 	echo "${resource_name/"node/"}"
 }
 
-# Deletes new_pod it wasn't present in the old_pods array.
-delete_pod_if_new() {
-	declare -r new_pod="$1"
-	shift
-	declare -r old_pods=("$@")
+# Get the new debugger pod that wasn't present in the old_pods array.
+get_new_debugger_pod() {
+    local old_pods=("$@")
+    local new_pod_list=($(kubectl get pods -o name | grep node-debugger))
 
-	for old_pod in "${old_pods[@]}"; do
-		[ "${old_pod}" == "${new_pod}" ] && return 0
-	done
-
-	kubectl delete "${new_pod}" >&2
+    for new_pod in "${new_pod_list[@]}"; do
+        if [[ ! " ${old_pods[*]} " =~ " ${new_pod} " ]]; then
+            echo "${new_pod}"
+            return
+        fi
+    done
 }
 
 # Runs a command in the host filesystem.
@@ -95,14 +95,24 @@ delete_pod_if_new() {
 #	$1 - the node name
 #
 exec_host() {
-	node="$1"
+	local node="$1"
 	# `kubectl debug` always returns 0, so we hack it to return the right exit code.
-	command="${@:2}"
+	local command="${@:2}"
 	command+='; echo -en \\n$?'
 
-	# Get the already existing debugger pods.
-	declare -a old_debugger_pods=( $(kubectl get pods -o name | grep node-debugger) )
+	# Get the already existing debugger pods
+	local old_debugger_pods=($(kubectl get pods -o name | grep node-debugger))
 
+	# Run a debug pod
+	kubectl debug -q "node/${node}" --image=quay.io/bedrock/ubuntu:latest -- chroot /host bash -c "sleep infinity" >&2
+
+	# Identify the new debugger pod
+	local new_debugger_pod=$(get_new_debugger_pod "${old_debugger_pods[@]}")
+
+	# Wait for the newly created pod to be ready
+	kubectl wait --timeout="30s" --for=condition=ready "${new_debugger_pod}" > /dev/null
+
+	# Execute the command and capture the output
 	# We're trailing the `\r` here due to: https://github.com/kata-containers/kata-containers/issues/8051
 	# tl;dr: When testing with CRI-O we're facing the following error:
 	# ```
@@ -112,17 +122,13 @@ exec_host() {
 	# [bats-exec-test:38] INFO: k8s configured to use runtimeclass
 	# bash: line 1: $'\r': command not found
 	# ```
-	output="$(kubectl debug -qi "node/${node}" --image=quay.io/bedrock/ubuntu:latest -- chroot /host bash -c "${command}" | tr -d '\r')"
+	local output="$(kubectl exec -qi "${new_debugger_pod}" -- chroot /host bash -c "${command}" | tr -d '\r')"
 
-	# Get the updated list of debugger pods.
-	declare -a new_debugger_pods=( $(kubectl get pods -o name | grep node-debugger) )
+	# Delete the newly created pod
+	kubectl delete "${new_debugger_pod}" >&2
 
-	# Delete the debugger pod created above.
-	for new_pod in "${new_debugger_pods[@]}"; do
-		delete_pod_if_new "${new_pod}" "${old_debugger_pods[@]}"
-	done
-
-	exit_code="$(echo "${output}" | tail -1)"
+	# Output the command result
+	local exit_code="$(echo "${output}" | tail -1)"
 	echo "$(echo "${output}" | head -n -1)"
 	return ${exit_code}
 }