5.3. 如何为应用程序创建健康检查脚本

工作负载健康检查脚本示例

# #!/bin/bash
set -e

SCRIPT_NAME=$(basename $0)
PODS_NS_LIST=(<user_workload_namespace1> <user_workload_namespace2>)
PODS_CT_LIST=(<user_workload_namespace1_pod_count> <user_workload_namespace2_pod_count>)
# Update these two lines with at least one namespace and the pod counts that are specific to your workloads. Use the kubernetes <namespace> where your workload is deployed.

# Set greenboot to read and execute the workload health check functions library.
source /usr/share/microshift/functions/greenboot.sh

# Set the exit handler to log the exit status.
trap 'script_exit' EXIT

# Set the script exit handler to log a `FAILURE` or `FINISHED` message depending on the exit status of the last command.
# args: None
# return: None
function script_exit() {
    [ "$?" -ne 0 ] && status=FAILURE || status=FINISHED
    echo $status
}

# Set the system to automatically stop the script if the user running it is not 'root'.
if [ $(id -u) -ne 0 ] ; then
    echo "The '${SCRIPT_NAME}' script must be run with the 'root' user privileges"
    exit 1
fi

echo "STARTED"

# Set the script to stop without reporting an error if the MicroShift service is not running.
if [ $(systemctl is-enabled microshift.service 2>/dev/null) != "enabled" ] ; then
    echo "MicroShift service is not enabled. Exiting..."
    exit 0
fi

# Set the wait timeout for the current check based on the boot counter.
WAIT_TIMEOUT_SECS=$(get_wait_timeout)

# Set the script to wait for the pod images to be downloaded.
for i in ${!PODS_NS_LIST[@]}; do
    CHECK_PODS_NS=${PODS_NS_LIST[$i]}

    echo "Waiting ${WAIT_TIMEOUT_SECS}s for pod image(s) from the ${CHECK_PODS_NS} namespace to be downloaded"
    wait_for ${WAIT_TIMEOUT_SECS} namespace_images_downloaded
done

# Set the script to wait for pods to enter ready state.
for i in ${!PODS_NS_LIST[@]}; do
    CHECK_PODS_NS=${PODS_NS_LIST[$i]}
    CHECK_PODS_CT=${PODS_CT_LIST[$i]}

    echo "Waiting ${WAIT_TIMEOUT_SECS}s for ${CHECK_PODS_CT} pod(s) from the ${CHECK_PODS_NS} namespace to be in 'Ready' state"
    wait_for ${WAIT_TIMEOUT_SECS} namespace_pods_ready
done

# Verify that pods are not restarting by running, which could indicate a crash loop.
for i in ${!PODS_NS_LIST[@]}; do
    CHECK_PODS_NS=${PODS_NS_LIST[$i]}

    echo "Checking pod restart count in the ${CHECK_PODS_NS} namespace"
    namespace_pods_not_restarting ${CHECK_PODS_NS}
done