Revert "fix cluster/log-dump/log-dump.sh shellcheck failures"

This reverts commit 742701622a.

It broke log dumping, see https://github.com/kubernetes/kubernetes/pull/88349#issuecomment-688385378
This commit is contained in:
Matt Matejczyk 2020-09-07 17:13:33 +02:00
parent b16070f3ce
commit 67f9d69eaf
2 changed files with 37 additions and 40 deletions

View File

@ -67,7 +67,7 @@ readonly max_dump_processes=25
function setup() {
KUBE_ROOT=$(dirname "${BASH_SOURCE[0]}")/../..
if [[ -z "${use_custom_instance_list}" ]]; then
: "${KUBE_CONFIG_FILE:="config-test.sh"}"
: ${KUBE_CONFIG_FILE:="config-test.sh"}
echo "Sourcing kube-util.sh"
source "${KUBE_ROOT}/cluster/kube-util.sh"
echo "Detecting project"
@ -89,7 +89,7 @@ function setup() {
}
function log-dump-ssh() {
if [[ "${gcloud_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
if [[ "${gcloud_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
ssh-to-node "$@"
return
fi
@ -106,7 +106,7 @@ function log-dump-ssh() {
function copy-logs-from-node() {
local -r node="${1}"
local -r dir="${2}"
while IFS= read -r line; do files+=("$line"); done <<< "$3"
local files=( ${3} )
# Append "*"
# The * at the end is needed to also copy rotated logs (which happens
# in large clusters and long runs).
@ -116,13 +116,12 @@ function copy-logs-from-node() {
# Comma delimit (even the singleton, or scp does the wrong thing), surround by braces.
local -r scp_files="{$(printf "%s," "${files[@]}")}"
if [[ "${gcloud_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
if [[ "${gcloud_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
# get-serial-port-output lets you ask for ports 1-4, but currently (11/21/2016) only port 1 contains useful information
gcloud compute instances get-serial-port-output --project "${PROJECT}" --zone "${ZONE}" --port 1 "${node}" > "${dir}/serial-1.log" || true
gcloud compute scp --recurse --project "${PROJECT}" --zone "${ZONE}" "${node}:${scp_files}" "${dir}" > /dev/null || true
elif [[ "${KUBERNETES_PROVIDER}" == "aws" ]]; then
local ip
ip=$(get_ssh_hostname "${node}")
local ip=$(get_ssh_hostname "${node}")
scp -oLogLevel=quiet -oConnectTimeout=30 -oStrictHostKeyChecking=no -i "${AWS_SSH_KEY}" "${SSH_USER}@${ip}:${scp_files}" "${dir}" > /dev/null || true
elif [[ -n "${use_custom_instance_list}" ]]; then
scp -oLogLevel=quiet -oConnectTimeout=30 -oStrictHostKeyChecking=no -i "${LOG_DUMP_SSH_KEY}" "${LOG_DUMP_SSH_USER}@${node}:${scp_files}" "${dir}" > /dev/null || true
@ -139,27 +138,26 @@ function copy-logs-from-node() {
function save-logs() {
local -r node_name="${1}"
local -r dir="${2}"
local save_files="${3}"
local files="${3}"
local opt_systemd_services="${4:-""}"
local on_master="${5:-"false"}"
save_files="${save_files} ${extra_log_files}"
files="${files} ${extra_log_files}"
if [[ -n "${use_custom_instance_list}" ]]; then
if [[ -n "${LOG_DUMP_SAVE_LOGS:-}" ]]; then
save_files="${save_files} ${LOG_DUMP_SAVE_LOGS:-}"
files="${files} ${LOG_DUMP_SAVE_LOGS:-}"
fi
else
case "${KUBERNETES_PROVIDER}" in
gce|gke)
save_files="${save_files} ${gce_logfiles}"
files="${files} ${gce_logfiles}"
;;
aws)
save_files="${save_files} ${aws_logfiles}"
files="${files} ${aws_logfiles}"
;;
esac
fi
read -r -a services <<< "${systemd_services} ${opt_systemd_services} ${extra_systemd_services}"
local -r services=( ${systemd_services} ${opt_systemd_services} ${extra_systemd_services} )
if log-dump-ssh "${node_name}" "command -v journalctl" &> /dev/null; then
if [[ "${on_master}" == "true" ]]; then
@ -179,7 +177,7 @@ function save-logs() {
log-dump-ssh "${node_name}" "sudo journalctl --output=short-precise" > "${dir}/systemd.log" || true
fi
else
save_files="${kern_logfile} ${save_files} ${initd_logfiles} ${supervisord_logfiles}"
files="${kern_logfile} ${files} ${initd_logfiles} ${supervisord_logfiles}"
fi
# Try dumping coverage profiles, if it looks like coverage is enabled in the first place.
@ -200,8 +198,8 @@ function save-logs() {
echo "Changing logfiles to be world-readable for download"
log-dump-ssh "${node_name}" "sudo chmod -R a+r /var/log" || true
echo "Copying '${save_files}' from ${node_name}"
copy-logs-from-node "${node_name}" "${dir}" "${save_files}"
echo "Copying '${files}' from ${node_name}"
copy-logs-from-node "${node_name}" "${dir}" "${files}"
}
# Saves a copy of the Windows Docker event log to ${WINDOWS_LOGS_DIR}\docker.log
@ -247,9 +245,8 @@ function save-windows-logs-via-diagnostics-tool() {
local node="${1}"
local dest_dir="${2}"
gcloud compute instances add-metadata "${node}" --metadata enable-diagnostics=true "--project=${PROJECT}" "--zone=${ZONE}"
local logs_archive_in_gcs
logs_archive_in_gcs=$(gcloud alpha compute diagnose export-logs "${node}" "--zone=${ZONE}" "--project=${PROJECT}" | tail -n 1)
gcloud compute instances add-metadata ${node} --metadata enable-diagnostics=true --project=${PROJECT} --zone=${ZONE}
local logs_archive_in_gcs=$(gcloud alpha compute diagnose export-logs ${node} --zone=${ZONE} --project=${PROJECT} | tail -n 1)
local temp_local_path="${node}.zip"
for retry in {1..20}; do
if gsutil mv "${logs_archive_in_gcs}" "${temp_local_path}" > /dev/null 2>&1; then
@ -261,8 +258,8 @@ function save-windows-logs-via-diagnostics-tool() {
done
if [[ -f "${temp_local_path}" ]]; then
unzip "${temp_local_path}" -d "${dest_dir}" > /dev/null
rm -f "${temp_local_path}"
unzip ${temp_local_path} -d "${dest_dir}" > /dev/null
rm -f ${temp_local_path}
fi
}
@ -275,14 +272,14 @@ function save-windows-logs-via-ssh() {
export-windows-docker-images-list "${node}"
local remote_files=()
for file in "${windows_node_logfiles[@]}"; do
for file in ${windows_node_logfiles[@]}; do
remote_files+=( "${WINDOWS_LOGS_DIR}\\${file}" )
done
remote_files+=( "${windows_node_otherfiles[@]}" )
# TODO(pjh, yujuhong): handle rotated logs and copying multiple files at the
# same time.
for remote_file in "${remote_files[@]}"; do
for remote_file in ${remote_files[@]}; do
# Retry up to 3 times to allow ssh keys to be properly propagated and
# stored.
for retry in {1..3}; do
@ -304,7 +301,7 @@ function save-logs-windows() {
local -r node="${1}"
local -r dest_dir="${2}"
if [[ ! "${gcloud_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
if [[ ! "${gcloud_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
echo "Not saving logs for ${node}, Windows log dumping requires gcloud support"
return
fi
@ -326,14 +323,14 @@ function run-in-docker-container() {
local node_name="$1"
local container="$2"
shift 2
log-dump-ssh "${node_name}" "docker exec \"\$(docker ps -f label=io.kubernetes.container.name=${container} --format \"{{.ID}}\")\" $*"
log-dump-ssh "${node_name}" "docker exec \"\$(docker ps -f label=io.kubernetes.container.name=${container} --format \"{{.ID}}\")\" $@"
}
function dump_masters() {
local master_names
if [[ -n "${use_custom_instance_list}" ]]; then
while IFS='' read -r line; do master_names+=("$line"); done < <(log_dump_custom_get_instances master)
elif [[ ! "${master_ssh_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
master_names=( $(log_dump_custom_get_instances master) )
elif [[ ! "${master_ssh_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
echo "Master SSH not supported for ${KUBERNETES_PROVIDER}"
return
elif [[ -n "${KUBEMARK_MASTER_NAME:-}" ]]; then
@ -384,8 +381,8 @@ function dump_nodes() {
node_names=( "$@" )
elif [[ -n "${use_custom_instance_list}" ]]; then
echo "Dumping logs for nodes provided by log_dump_custom_get_instances() function"
while IFS='' read -r line; do node_names+=("$line"); done < <(log_dump_custom_get_instances node)
elif [[ ! "${node_ssh_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
node_names=( $(log_dump_custom_get_instances node) )
elif [[ ! "${node_ssh_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
echo "Node SSH not supported for ${KUBERNETES_PROVIDER}"
return
else
@ -412,7 +409,7 @@ function dump_nodes() {
linux_nodes_selected_for_logs=()
if [[ -n "${LOGDUMP_ONLY_N_RANDOM_NODES:-}" ]]; then
# We randomly choose 'LOGDUMP_ONLY_N_RANDOM_NODES' many nodes for fetching logs.
for index in $(shuf -i 0-$(( ${#node_names[*]} - 1 )) -n "${LOGDUMP_ONLY_N_RANDOM_NODES}")
for index in `shuf -i 0-$(( ${#node_names[*]} - 1 )) -n ${LOGDUMP_ONLY_N_RANDOM_NODES}`
do
linux_nodes_selected_for_logs+=("${node_names[$index]}")
done
@ -478,7 +475,7 @@ function find_non_logexported_nodes() {
echo "Successfully listed marker files for successful nodes"
NON_LOGEXPORTED_NODES=()
for node in "${NODE_NAMES[@]}"; do
if [[ ! "${succeeded_nodes}" =~ ${node} ]]; then
if [[ ! "${succeeded_nodes}" =~ "${node}" ]]; then
NON_LOGEXPORTED_NODES+=("${node}")
fi
done
@ -489,8 +486,7 @@ function find_non_logexported_nodes() {
function dump_nodes_with_logexporter() {
if [[ -n "${use_custom_instance_list}" ]]; then
echo "Dumping logs for nodes provided by log_dump_custom_get_instances() function"
NODE_NAMES=()
while IFS='' read -r line; do NODE_NAMES+=("$line"); done < <(log_dump_custom_get_instances node)
NODE_NAMES=( $(log_dump_custom_get_instances node) )
else
echo "Detecting nodes in the cluster"
detect-node-names &> /dev/null
@ -502,7 +498,7 @@ function dump_nodes_with_logexporter() {
fi
# Obtain parameters required by logexporter.
local -r service_account_credentials="$(base64 <<< "${GOOGLE_APPLICATION_CREDENTIALS}" | tr -d '\n')"
local -r service_account_credentials="$(cat ${GOOGLE_APPLICATION_CREDENTIALS} | base64 | tr -d '\n')"
local -r cloud_provider="${KUBERNETES_PROVIDER}"
local -r enable_hollow_node_logs="${ENABLE_HOLLOW_NODE_LOGS:-false}"
local -r logexport_sleep_seconds="$(( 90 + NUM_NODES / 3 ))"
@ -563,6 +559,7 @@ function dump_nodes_with_logexporter() {
done; wait)
# List registry of marker files (of nodes whose logexporter succeeded) from GCS.
local nodes_succeeded
for retry in {1..10}; do
if find_non_logexported_nodes; then
break
@ -592,29 +589,28 @@ function dump_nodes_with_logexporter() {
"${KUBECTL}" get pods --namespace "${logexporter_namespace}" || true
"${KUBECTL}" delete namespace "${logexporter_namespace}" || true
if [[ "${#failed_nodes[@]}" != 0 ]]; then
echo -e "Dumping logs through SSH for the following nodes:\n${failed_nodes[*]}"
echo -e "Dumping logs through SSH for the following nodes:\n${failed_nodes[@]}"
dump_nodes "${failed_nodes[@]}"
fi
}
function detect_node_failures() {
if ! [[ "${gcloud_supported_providers}" =~ ${KUBERNETES_PROVIDER} ]]; then
if ! [[ "${gcloud_supported_providers}" =~ "${KUBERNETES_PROVIDER}" ]]; then
return
fi
detect-node-names
if [[ "${KUBERNETES_PROVIDER}" == "gce" ]]; then
local all_instance_groups=("${INSTANCE_GROUPS[@]}" "${WINDOWS_INSTANCE_GROUPS[@]}")
local all_instance_groups=(${INSTANCE_GROUPS[@]} ${WINDOWS_INSTANCE_GROUPS[@]})
else
local all_instance_groups=("${INSTANCE_GROUPS[@]}")
local all_instance_groups=(${INSTANCE_GROUPS[@]})
fi
if [ -z "${all_instance_groups:-}" ]; then
return
fi
for group in "${all_instance_groups[@]}"; do
local creation_timestamp
creation_timestamp=$(gcloud compute instance-groups managed describe \
local creation_timestamp=$(gcloud compute instance-groups managed describe \
"${group}" \
--project "${PROJECT}" \
--zone "${ZONE}" \

View File

@ -1,3 +1,4 @@
./cluster/gce/gci/configure.sh
./cluster/gce/gci/master-helper.sh
./cluster/gce/util.sh
./cluster/log-dump/log-dump.sh