1
0
mirror of https://github.com/mailcow/mailcow-dockerized.git synced 2024-12-14 10:52:49 +02:00
mailcow-dockerized/data/Dockerfiles/watchdog/watchdog.sh

646 lines
23 KiB
Bash
Raw Normal View History

2017-09-20 10:56:49 +02:00
#!/bin/bash
trap "exit" INT TERM
trap "kill 0" EXIT
# Prepare
BACKGROUND_TASKS=()
2019-03-28 00:15:04 +02:00
echo "Waiting for containers to settle..."
sleep 10
2017-09-20 10:56:49 +02:00
2017-10-03 12:05:38 +02:00
if [[ "${USE_WATCHDOG}" =~ ^([nN][oO]|[nN])+$ ]]; then
2017-10-03 12:07:48 +02:00
echo -e "$(date) - USE_WATCHDOG=n, skipping watchdog..."
2017-10-03 12:05:38 +02:00
sleep 365d
exec $(readlink -f "$0")
fi
2017-09-20 10:56:49 +02:00
# Checks pipe their corresponding container name in this pipe
if [[ ! -p /tmp/com_pipe ]]; then
mkfifo /tmp/com_pipe
fi
# Common functions
progress() {
SERVICE=${1}
TOTAL=${2}
CURRENT=${3}
DIFF=${4}
[[ -z ${DIFF} ]] && DIFF=0
[[ -z ${TOTAL} || -z ${CURRENT} ]] && return
[[ ${CURRENT} -gt ${TOTAL} ]] && return
[[ ${CURRENT} -lt 0 ]] && CURRENT=0
2017-09-20 23:24:56 +02:00
PERCENT=$(( 200 * ${CURRENT} / ${TOTAL} % 2 + 100 * ${CURRENT} / ${TOTAL} ))
2017-12-09 14:15:24 +02:00
redis-cli -h redis LPUSH WATCHDOG_LOG "{\"time\":\"$(date +%s)\",\"service\":\"${SERVICE}\",\"lvl\":\"${PERCENT}\",\"hpnow\":\"${CURRENT}\",\"hptotal\":\"${TOTAL}\",\"hpdiff\":\"${DIFF}\"}" > /dev/null
log_msg "${SERVICE} health level: ${PERCENT}% (${CURRENT}/${TOTAL}), health trend: ${DIFF}" no_redis
# Return 10 to indicate a dead service
[ ${CURRENT} -le 0 ] && return 10
2017-09-20 10:56:49 +02:00
}
log_msg() {
2017-12-09 14:15:24 +02:00
if [[ ${2} != "no_redis" ]]; then
redis-cli -h redis LPUSH WATCHDOG_LOG "{\"time\":\"$(date +%s)\",\"message\":\"$(printf '%s' "${1}" | \
tr '\r\n%&;$"_[]{}-' ' ')\"}" > /dev/null
2017-12-09 14:15:24 +02:00
fi
echo $(date) $(printf '%s\n' "${1}")
}
function mail_error() {
[[ -z ${1} ]] && return 1
[[ -z ${2} ]] && BODY="Service was restarted on $(date), please check your mailcow installation." || BODY="$(date) - ${2}"
WATCHDOG_NOTIFY_EMAIL=$(echo "${WATCHDOG_NOTIFY_EMAIL}" | sed 's/"//;s|"$||')
IFS=',' read -r -a MAIL_RCPTS <<< "${WATCHDOG_NOTIFY_EMAIL}"
for rcpt in "${MAIL_RCPTS[@]}"; do
RCPT_DOMAIN=
RCPT_MX=
RCPT_DOMAIN=$(echo ${rcpt} | awk -F @ {'print $NF'})
RCPT_MX=$(dig +short ${RCPT_DOMAIN} mx | sort -n | awk '{print $2; exit}')
if [[ -z ${RCPT_MX} ]]; then
log_msg "Cannot determine MX for ${rcpt}, skipping email notification..."
return 1
fi
[ -f "/tmp/${1}" ] && ATTACH="--attach /tmp/${1}@text/plain" || ATTACH=
./smtp-cli --missing-modules-ok \
--subject="Watchdog: ${1} hit the error rate limit" \
--body-plain="${BODY}" \
--to=${rcpt} \
--from="watchdog@${MAILCOW_HOSTNAME}" \
--server="${RCPT_MX}" \
--hello-host=${MAILCOW_HOSTNAME} \
${ATTACH}
log_msg "Sent notification email to ${rcpt}"
done
}
2017-09-20 12:27:24 +02:00
get_container_ip() {
# ${1} is container
CONTAINER_ID=()
CONTAINER_IPS=()
2017-09-20 12:27:24 +02:00
CONTAINER_IP=
2017-09-20 23:24:56 +02:00
LOOP_C=1
until [[ ${CONTAINER_IP} =~ ^[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}$ ]] || [[ ${LOOP_C} -gt 5 ]]; do
if [ ${IP_BY_DOCKER_API} -eq 0 ]; then
CONTAINER_IP=$(dig a "${1}" +short)
else
sleep 0.5
# get long container id for exact match
CONTAINER_ID=($(curl --silent --insecure https://dockerapi/containers/json | jq -r ".[] | {name: .Config.Labels[\"com.docker.compose.service\"], id: .Id}" | jq -rc "select( .name | tostring == \"${1}\") | .id"))
# returned id can have multiple elements (if scaled), shuffle for random test
CONTAINER_ID=($(printf "%s\n" "${CONTAINER_ID[@]}" | shuf))
if [[ ! -z ${CONTAINER_ID} ]]; then
for matched_container in "${CONTAINER_ID[@]}"; do
CONTAINER_IPS=($(curl --silent --insecure https://dockerapi/containers/${matched_container}/json | jq -r '.NetworkSettings.Networks[].IPAddress'))
for ip_match in "${CONTAINER_IPS[@]}"; do
# grep will do nothing if one of these vars is empty
[[ -z ${ip_match} ]] && continue
[[ -z ${IPV4_NETWORK} ]] && continue
# only return ips that are part of our network
if ! grep -q ${IPV4_NETWORK} <(echo ${ip_match}); then
continue
else
CONTAINER_IP=${ip_match}
break
fi
done
[[ ! -z ${CONTAINER_IP} ]] && break
done
fi
fi
2017-09-20 23:24:56 +02:00
LOOP_C=$((LOOP_C + 1))
2017-09-20 12:27:24 +02:00
done
2017-09-20 23:24:56 +02:00
[[ ${LOOP_C} -gt 5 ]] && echo 240.0.0.0 || echo ${CONTAINER_IP}
2017-09-20 12:27:24 +02:00
}
2017-09-20 10:56:49 +02:00
nginx_checks() {
err_count=0
diff_c=0
THRESHOLD=16
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/nginx-mailcow; echo "$(tail -50 /tmp/nginx-mailcow)" > /tmp/nginx-mailcow
2017-09-20 23:24:56 +02:00
host_ip=$(get_container_ip nginx-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_http -4 -H ${host_ip} -u / -p 8081 2>> /tmp/nginx-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Nginx" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
unbound_checks() {
err_count=0
diff_c=0
THRESHOLD=8
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/unbound-mailcow; echo "$(tail -50 /tmp/unbound-mailcow)" > /tmp/unbound-mailcow
host_ip=$(get_container_ip unbound-mailcow)
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_dns -s ${host_ip} -H stackoverflow.com 2>> /tmp/unbound-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
DNSSEC=$(dig com +dnssec | egrep 'flags:.+ad')
if [[ -z ${DNSSEC} ]]; then
echo "DNSSEC failure" 2>> /tmp/unbound-mailcow 1>&2
err_count=$(( ${err_count} + 1))
else
echo "DNSSEC check succeeded" 2>> /tmp/unbound-mailcow 1>&2
fi
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Unbound" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
done
return 1
}
2017-09-20 10:56:49 +02:00
mysql_checks() {
err_count=0
diff_c=0
THRESHOLD=12
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/mysql-mailcow; echo "$(tail -50 /tmp/mysql-mailcow)" > /tmp/mysql-mailcow
2017-09-20 23:24:56 +02:00
host_ip=$(get_container_ip mysql-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_mysql -s /var/run/mysqld/mysqld.sock -u ${DBUSER} -p ${DBPASS} -d ${DBNAME} 2>> /tmp/mysql-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_mysql_query -s /var/run/mysqld/mysqld.sock -u ${DBUSER} -p ${DBPASS} -d ${DBNAME} -q "SELECT COUNT(*) FROM information_schema.tables" 2>> /tmp/mysql-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "MySQL/MariaDB" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
sogo_checks() {
err_count=0
diff_c=0
THRESHOLD=10
2017-09-20 10:56:49 +02:00
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/sogo-mailcow; echo "$(tail -50 /tmp/sogo-mailcow)" > /tmp/sogo-mailcow
2017-09-20 23:24:56 +02:00
host_ip=$(get_container_ip sogo-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_http -4 -H ${host_ip} -u /SOGo.index/ -p 20000 -R "SOGo\.MainUI" 2>> /tmp/sogo-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "SOGo" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
postfix_checks() {
err_count=0
diff_c=0
THRESHOLD=8
2017-09-20 10:56:49 +02:00
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/postfix-mailcow; echo "$(tail -50 /tmp/postfix-mailcow)" > /tmp/postfix-mailcow
host_ip=$(get_container_ip postfix-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_smtp -4 -H ${host_ip} -p 589 -f "watchdog@invalid" -C "RCPT TO:null@localhost" -C DATA -C . -R 250 2>> /tmp/postfix-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_smtp -4 -H ${host_ip} -p 589 -S 2>> /tmp/postfix-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Postfix" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
clamd_checks() {
err_count=0
diff_c=0
THRESHOLD=15
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/clamd-mailcow; echo "$(tail -50 /tmp/clamd-mailcow)" > /tmp/clamd-mailcow
host_ip=$(get_container_ip clamd-mailcow)
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_clamd -4 -H ${host_ip} 2>> /tmp/clamd-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Clamd" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 30 ))
fi
done
return 1
}
2017-09-20 10:56:49 +02:00
dovecot_checks() {
err_count=0
diff_c=0
THRESHOLD=20
2017-09-20 10:56:49 +02:00
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/dovecot-mailcow; echo "$(tail -50 /tmp/dovecot-mailcow)" > /tmp/dovecot-mailcow
2017-09-20 23:24:56 +02:00
host_ip=$(get_container_ip dovecot-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_smtp -4 -H ${host_ip} -p 24 -f "watchdog@invalid" -C "RCPT TO:<watchdog@invalid>" -L -R "User doesn't exist" 2>> /tmp/dovecot-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_imap -4 -H ${host_ip} -p 993 -S -e "OK " 2>> /tmp/dovecot-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_imap -4 -H ${host_ip} -p 143 -e "OK " 2>> /tmp/dovecot-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_tcp -4 -H ${host_ip} -p 10001 -e "VERSION" 2>> /tmp/dovecot-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_tcp -4 -H ${host_ip} -p 4190 -e "Dovecot ready" 2>> /tmp/dovecot-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Dovecot" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
phpfpm_checks() {
err_count=0
diff_c=0
THRESHOLD=5
2017-09-20 10:56:49 +02:00
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/php-fpm-mailcow; echo "$(tail -50 /tmp/php-fpm-mailcow)" > /tmp/php-fpm-mailcow
2017-09-20 23:24:56 +02:00
host_ip=$(get_container_ip php-fpm-mailcow)
2017-09-20 10:56:49 +02:00
err_c_cur=${err_count}
/usr/lib/nagios/plugins/check_tcp -H ${host_ip} -p 9001 2>> /tmp/php-fpm-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
/usr/lib/nagios/plugins/check_tcp -H ${host_ip} -p 9002 2>> /tmp/php-fpm-mailcow 1>&2; err_count=$(( ${err_count} + $? ))
2017-09-20 10:56:49 +02:00
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "PHP-FPM" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
2017-09-20 10:56:49 +02:00
done
return 1
}
ratelimit_checks() {
err_count=0
diff_c=0
THRESHOLD=1
RL_LOG_STATUS=$(redis-cli -h redis LRANGE RL_LOG 0 0 | jq .qid)
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
err_c_cur=${err_count}
RL_LOG_STATUS_PREV=${RL_LOG_STATUS}
RL_LOG_STATUS=$(redis-cli -h redis LRANGE RL_LOG 0 0 | jq .qid)
if [[ ${RL_LOG_STATUS_PREV} != ${RL_LOG_STATUS} ]]; then
err_count=$(( ${err_count} + 1 ))
fi
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Ratelimit" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
done
return 1
}
2019-03-28 00:15:04 +02:00
acme_checks() {
err_count=0
diff_c=0
THRESHOLD=1
ACME_LOG_STATUS=$(redis-cli -h redis GET ACME_FAIL_TIME)
if [[ -z "${ACME_LOG_STATUS}" ]]; then
redis-cli -h redis SET ACME_FAIL_TIME 0
ACME_LOG_STATUS=0
fi
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
err_c_cur=${err_count}
ACME_LOG_STATUS_PREV=${ACME_LOG_STATUS}
ACME_LOG_STATUS=$(redis-cli -h redis GET ACME_FAIL_TIME)
if [[ ${ACME_LOG_STATUS_PREV} != ${ACME_LOG_STATUS} ]]; then
err_count=$(( ${err_count} + 1 ))
fi
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "ACME" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
fi
done
return 1
}
ipv6nat_checks() {
err_count=0
diff_c=0
THRESHOLD=1
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
err_c_cur=${err_count}
CONTAINERS=$(curl --silent --insecure https://dockerapi/containers/json)
IPV6NAT_CONTAINER_ID=$(echo ${CONTAINERS} | jq -r ".[] | {name: .Config.Labels[\"com.docker.compose.service\"], id: .Id}" | jq -rc "select( .name | tostring | contains(\"ipv6nat-mailcow\")) | .id")
if [[ ! -z ${IPV6NAT_CONTAINER_ID} ]]; then
LATEST_STARTED="$(echo ${CONTAINERS} | jq -r ".[] | {name: .Config.Labels[\"com.docker.compose.service\"], StartedAt: .State.StartedAt}" | jq -rc "select( .name | tostring | contains(\"ipv6nat-mailcow\") | not)" | jq -rc .StartedAt | xargs -n1 date +%s -d | sort | tail -n1)"
LATEST_IPV6NAT="$(echo ${CONTAINERS} | jq -r ".[] | {name: .Config.Labels[\"com.docker.compose.service\"], StartedAt: .State.StartedAt}" | jq -rc "select( .name | tostring | contains(\"ipv6nat-mailcow\"))" | jq -rc .StartedAt | xargs -n1 date +%s -d | sort | tail -n1)"
DIFFERENCE_START_TIME=$(expr ${LATEST_IPV6NAT} - ${LATEST_STARTED} 2>/dev/null)
if [[ "${DIFFERENCE_START_TIME}" -lt 30 ]]; then
err_count=$(( ${err_count} + 1 ))
fi
fi
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "IPv6 NAT" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
if [[ $? == 10 ]]; then
diff_c=0
sleep 1
else
diff_c=0
sleep 300
fi
done
return 1
}
rspamd_checks() {
err_count=0
diff_c=0
THRESHOLD=5
# Reduce error count by 2 after restarting an unhealthy container
trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
while [ ${err_count} -lt ${THRESHOLD} ]; do
touch /tmp/rspamd-mailcow; echo "$(tail -50 /tmp/rspamd-mailcow)" > /tmp/rspamd-mailcow
host_ip=$(get_container_ip rspamd-mailcow)
err_c_cur=${err_count}
SCORE=$(echo 'To: null@localhost
From: watchdog@localhost
Empty
' | usr/bin/curl -s --data-binary @- --unix-socket /var/lib/rspamd/rspamd.sock http://rspamd/scan | jq -rc .required_score)
if [[ ${SCORE} != "9999" ]]; then
echo "Rspamd settings check failed" 2>> /tmp/rspamd-mailcow 1>&2
err_count=$(( ${err_count} + 1))
else
echo "Rspamd settings check succeeded" 2>> /tmp/rspamd-mailcow 1>&2
fi
[ ${err_c_cur} -eq ${err_count} ] && [ ! $((${err_count} - 1)) -lt 0 ] && err_count=$((${err_count} - 1)) diff_c=1
[ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
progress "Rspamd" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
2017-09-22 16:40:02 +02:00
diff_c=0
sleep $(( ( RANDOM % 30 ) + 10 ))
done
return 1
}
2017-09-20 10:56:49 +02:00
# Create watchdog agents
(
while true; do
if ! nginx_checks; then
log_msg "Nginx hit error limit"
2017-09-20 10:56:49 +02:00
echo nginx-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! mysql_checks; then
log_msg "MySQL hit error limit"
2017-09-20 10:56:49 +02:00
echo mysql-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! phpfpm_checks; then
log_msg "PHP-FPM hit error limit"
2017-09-20 10:56:49 +02:00
echo php-fpm-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! sogo_checks; then
log_msg "SOGo hit error limit"
2017-09-20 10:56:49 +02:00
echo sogo-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
if [ ${CHECK_UNBOUND} -eq 1 ]; then
(
while true; do
if ! unbound_checks; then
log_msg "Unbound hit error limit"
echo unbound-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
fi
if [[ "${SKIP_CLAMD}" =~ ^([nN][oO]|[nN])+$ ]]; then
(
while true; do
if ! clamd_checks; then
log_msg "Clamd hit error limit"
echo clamd-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
fi
2017-09-20 10:56:49 +02:00
(
while true; do
if ! postfix_checks; then
log_msg "Postfix hit error limit"
2017-09-20 10:56:49 +02:00
echo postfix-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! dovecot_checks; then
log_msg "Dovecot hit error limit"
2017-09-20 10:56:49 +02:00
echo dovecot-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! rspamd_checks; then
log_msg "Rspamd hit error limit"
echo rspamd-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! ratelimit_checks; then
log_msg "Ratelimit hit error limit"
echo ratelimit > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
2019-03-28 00:15:04 +02:00
(
while true; do
if ! acme_checks; then
log_msg "ACME client hit error limit"
echo acme-tiny > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
(
while true; do
if ! ipv6nat_checks; then
log_msg "IPv6 NAT warning: ipv6nat-mailcow container was not started at least 30s after siblings (not an error)"
echo ipv6nat-mailcow > /tmp/com_pipe
fi
done
) &
BACKGROUND_TASKS+=($!)
2017-09-20 10:56:49 +02:00
# Monitor watchdog agents, stop script when agents fails and wait for respawn by Docker (restart:always:n)
(
while true; do
for bg_task in ${BACKGROUND_TASKS[*]}; do
if ! kill -0 ${bg_task} 1>&2; then
log_msg "Worker ${bg_task} died, stopping watchdog and waiting for respawn..."
kill -TERM 1
2017-09-20 10:56:49 +02:00
fi
sleep 10
2017-09-20 10:56:49 +02:00
done
done
) &
# Monitor dockerapi
(
while true; do
while nc -z dockerapi 443; do
sleep 3
done
log_msg "Cannot find dockerapi-mailcow, waiting to recover..."
kill -STOP ${BACKGROUND_TASKS[*]}
until nc -z dockerapi 443; do
sleep 3
done
kill -CONT ${BACKGROUND_TASKS[*]}
kill -USR1 ${BACKGROUND_TASKS[*]}
done
) &
2017-09-20 10:56:49 +02:00
# Restart container when threshold limit reached
while true; do
CONTAINER_ID=
HAS_INITDB=
2017-09-20 10:56:49 +02:00
read com_pipe_answer </tmp/com_pipe
if [ -s "/tmp/${com_pipe_answer}" ]; then
cat "/tmp/${com_pipe_answer}"
fi
if [[ ${com_pipe_answer} == "ratelimit" ]]; then
log_msg "At least one ratelimit was applied"
2019-03-28 00:15:04 +02:00
[[ ! -z ${WATCHDOG_NOTIFY_EMAIL} ]] && mail_error "${com_pipe_answer}" "Please see mailcow UI logs for further information."
elif [[ ${com_pipe_answer} == "acme-tiny" ]]; then
log_msg "acme-tiny client returned non-zero exit code"
[[ ! -z ${WATCHDOG_NOTIFY_EMAIL} ]] && mail_error "${com_pipe_answer}" "Please check acme-mailcow for ruther information."
elif [[ ${com_pipe_answer} =~ .+-mailcow ]] || [[ ${com_pipe_answer} == "ipv6nat-mailcow" ]]; then
2017-09-20 10:56:49 +02:00
kill -STOP ${BACKGROUND_TASKS[*]}
sleep 3
CONTAINER_ID=$(curl --silent --insecure https://dockerapi/containers/json | jq -r ".[] | {name: .Config.Labels[\"com.docker.compose.service\"], id: .Id}" | jq -rc "select( .name | tostring | contains(\"${com_pipe_answer}\")) | .id")
2017-09-20 10:56:49 +02:00
if [[ ! -z ${CONTAINER_ID} ]]; then
if [[ "${com_pipe_answer}" == "php-fpm-mailcow" ]]; then
HAS_INITDB=$(curl --silent --insecure -XPOST https://dockerapi/containers/${CONTAINER_ID}/top | jq '.msg.Processes[] | contains(["php -c /usr/local/etc/php -f /web/inc/init_db.inc.php"])' | grep true)
fi
S_RUNNING=$(($(date +%s) - $(curl --silent --insecure https://dockerapi/containers/${CONTAINER_ID}/json | jq .State.StartedAt | xargs -n1 date +%s -d)))
if [ ${S_RUNNING} -lt 120 ]; then
log_msg "Container is running for less than 120 seconds, skipping action..."
elif [[ ! -z ${HAS_INITDB} ]]; then
log_msg "Database is being initialized by php-fpm-mailcow, not restarting but delaying checks for a minute..."
sleep 60
else
log_msg "Sending restart command to ${CONTAINER_ID}..."
curl --silent --insecure -XPOST https://dockerapi/containers/${CONTAINER_ID}/restart
if [[ ${com_pipe_answer} != "ipv6nat-mailcow" ]]; then
[[ ! -z ${WATCHDOG_NOTIFY_EMAIL} ]] && mail_error "${com_pipe_answer}"
fi
log_msg "Wait for restarted container to settle and continue watching..."
sleep 35
fi
2017-09-20 10:56:49 +02:00
fi
kill -CONT ${BACKGROUND_TASKS[*]}
kill -USR1 ${BACKGROUND_TASKS[*]}
fi
done