[CI] Improve logging in CI scripts and add timeouts (#195)

- show events every 15 seconds
- dump logs every 5 minutes
- dump logs when timeouting
This commit is contained in:
Lari Hotari 2022-01-04 17:24:28 +02:00 committed by GitHub
parent 9b672379dd
commit d3589b1c82
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -29,6 +29,28 @@ NAMESPACE=pulsar
CLUSTER=pulsar-ci
CLUSTER_ID=$(uuidgen)
# brew package 'coreutils' is required on MacOSX
# coreutils includes the 'timeout' command
if [[ "$OSTYPE" == "darwin"* ]]; then
brew_gnubin_packages=(coreutils)
if ! type -P brew &>/dev/null; then
echo "On MacOSX, you must install required binaries with the following command:"
echo "brew install" "${brew_gnubin_packages[@]}"
exit 1
fi
for dep in "${brew_gnubin_packages[@]}"; do
path_element="$(brew --prefix)/opt/${dep}/libexec/gnubin"
if [ ! -d "${path_element}" ]; then
echo "'${path_element}' is missing. Quick fix: 'brew install ${dep}'."
echo "On MacOSX, you must install required binaries with the following command:"
echo "brew install" "${brew_gnubin_packages[@]}"
exit 1
fi
PATH="${path_element}:$PATH"
done
export PATH
fi
function ci::create_cluster() {
echo "Creating a kind cluster ..."
${CHARTS_HOME}/hack/kind-cluster-build.sh --name pulsar-ci-${CLUSTER_ID} -c 1 -v 10
@ -65,11 +87,19 @@ function ci::install_cert_manager() {
echo ${WC};
sleep 15
${KUBECTL} get pods -n cert-manager
${KUBECTL} get events --sort-by=.lastTimestamp -A | tail -n 30 || true
WC=$(${KUBECTL} get pods -n cert-manager --field-selector=status.phase=Running | wc -l)
done
echo "Successfully installed the cert manager."
}
function ci::print_pod_logs() {
echo "Logs for all pulsar containers:"
for pod in $(${KUBECTL} get pods -n ${NAMESPACE} -l app=pulsar -o=name); do
${KUBECTL} logs -n ${NAMESPACE} "$pod" --all-containers=true --ignore-errors=true --prefix=true --tail=100 || true
done;
}
function ci::install_pulsar_chart() {
local value_file=$1
local extra_opts=$2
@ -87,10 +117,20 @@ function ci::install_pulsar_chart() {
echo "wait until broker is alive"
WC=$(${KUBECTL} get pods -n ${NAMESPACE} --field-selector=status.phase=Running | grep ${CLUSTER}-broker | wc -l)
counter=1
while [[ ${WC} -lt 1 ]]; do
((counter++))
echo ${WC};
sleep 15
${KUBECTL} get pods -n ${NAMESPACE}
${KUBECTL} get events --sort-by=.lastTimestamp -A | tail -n 30 || true
if [[ $((counter % 20)) -eq 0 ]]; then
ci::print_pod_logs
if [[ $counter -gt 100 ]]; then
echo >&2 "Timeout waiting..."
exit 1
fi
fi
WC=$(${KUBECTL} get pods -n ${NAMESPACE} | grep ${CLUSTER}-broker | wc -l)
if [[ ${WC} -gt 1 ]]; then
${KUBECTL} describe pod -n ${NAMESPACE} pulsar-ci-broker-0
@ -98,17 +138,27 @@ function ci::install_pulsar_chart() {
fi
WC=$(${KUBECTL} get pods -n ${NAMESPACE} --field-selector=status.phase=Running | grep ${CLUSTER}-broker | wc -l)
done
${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until nslookup pulsar-ci-broker; do sleep 3; done'
${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until [ "$(curl -L http://pulsar-ci-broker:8080/status.html)" == "OK" ]; do sleep 3; done'
timeout 300s ${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until nslookup pulsar-ci-broker; do sleep 3; done' || { echo >&2 "Timeout waiting..."; ci::print_pod_logs; exit 1; }
timeout 120s ${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until [ "$(curl -L http://pulsar-ci-broker:8080/status.html)" == "OK" ]; do sleep 3; done' || { echo >&2 "Timeout waiting..."; ci::print_pod_logs; exit 1; }
WC=$(${KUBECTL} get pods -n ${NAMESPACE} --field-selector=status.phase=Running | grep ${CLUSTER}-proxy | wc -l)
counter=1
while [[ ${WC} -lt 1 ]]; do
((counter++))
echo ${WC};
sleep 15
${KUBECTL} get pods -n ${NAMESPACE}
${KUBECTL} get events --sort-by=.lastTimestamp -A | tail -n 30 || true
if [[ $((counter % 8)) -eq 0 ]]; then
ci::print_pod_logs
if [[ $counter -gt 16 ]]; then
echo >&2 "Timeout waiting..."
exit 1
fi
fi
WC=$(${KUBECTL} get pods -n ${NAMESPACE} --field-selector=status.phase=Running | grep ${CLUSTER}-proxy | wc -l)
done
${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until nslookup pulsar-ci-proxy; do sleep 3; done'
timeout 300s ${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until nslookup pulsar-ci-proxy; do sleep 3; done' || { echo >&2 "Timeout waiting..."; ci::print_pod_logs; exit 1; }
# ${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'until [ "$(curl -L http://pulsar-ci-proxy:8080/status.html)" == "OK" ]; do sleep 3; done'
}
@ -131,6 +181,7 @@ function ci::wait_function_running() {
echo ${num_running}
sleep 15
${KUBECTL} get pods -n ${NAMESPACE} --field-selector=status.phase=Running
${KUBECTL} get events --sort-by=.lastTimestamp -A | tail -n 30 || true
num_running=$(${KUBECTL} exec -n ${NAMESPACE} ${CLUSTER}-toolset-0 -- bash -c 'bin/pulsar-admin functions status --tenant pulsar-ci --namespace test --name test-function | bin/jq .numRunning')
done
}