badi/public_scripts/canardien/canardien - view

File: [Local Repository] / badi / public_scripts / canardien / canardien
Revision 1.6: download - view: text, annotated - select for diffs - revision graph
Sun Aug 6 20:06:51 2023 UTC (10 months, 1 week ago) by adi
Branches: MAIN
CVS tags: HEAD

Parallel execution when guarding multiple hosts.

1: #!/bin/bash 2: 3: # canardien 0.0.6 4: # (c) 2005-2021 under GPL by Adrian Zaugg 5: 6: 7: # canardien [<host> [<host> ...]] 8: 9: # canardiens pings a machine, originally ente.limmat.ch thus its name, 10: # to determine wether she is gone diving. 11: 12: 13: ## Settings 14: # 15: 16: # Hosts to ping. use a space separated list for multiple targets. All 17: # hosts of the list are checked and their results reported individually. 18: # Hosts given on the command line overwrite this setting. 19: HOSTS="" 20: 21: # eMail Alerts 22: # Send alert email messages to the following address(es). Leave empty 23: # for no alert. For multiple destinations use a comma separated list. 24: ALERT_TO="root" 25: 26: # Subject of alert email 27: ALERT_SUBJECT='"Attention: no answer from $HOST anymore!"' 28: 29: # Text of Message (Put variables in 'single quotes' to protect them. They 30: # should get expanded at the time the message is sent!) 31: ALERT_TEXT='"\n[$TIME_STAMP]\n\nALERT!!\n\n\t$HOST is down!\n\nYou should probably do something, please.\n\nKind regards, $PINGHOST."' 32: 33: # Path to fping 34: PING="" 35: 36: # Answer of fping to reachable hosts 37: ALIVE_ANSWER="is alive" 38: 39: # Max number of pakets to send before giving up. Time increases exponentially, 40: # use a number < 7. 41: RETRIES=5 42: 43: # Temporary file path 44: TMPDIR="/tmp" 45: 46: # This hosts name 47: PINGHOST="$(uname -n)" 48: 49: # Set to an empty string to avoid debug output, 50: # to "low" for a few, output and to anything else 51: # for verbose output 52: DEBUG=verbose 53: 54: # For silent (non-error) operation set to anything, 55: # comment out to enable text output 56: #SILENT=shshsh 57: 58: 59: # -----------functions----------- 60: 61: # Initialize. 62: function init() { 63: 64: # check fping existence 65: if [ -z "$PING" ]; then 66: PING="$(which fping)" 67: fi 68: if [ ${#PING} -eq 0 ]; then 69: echo "Error: fping external program not found or not set. Exitting." >&2 70: exit 1 71: fi 72: if [ ! -x "$PING" ]; then 73: echo "Error: Can't execute the program set to use as fping. Please enter the correct path to \"fping\"." >&2 74: exit 1 75: fi 76: 77: # get hostset 78: if [ -n "$*" ]; then 79: HOSTS="$*" 80: fi 81: 82: # remove multiple spaces and trim 83: HOSTS="$(echo "$HOSTS" | sed -e "s/ \{1,\}/ /g" -e "s/^ //" -e "s/ $//")" 84: 85: if [ -z "$HOSTS" ]; then 86: echo "Error: No host to ping." >&2 87: exit 1 88: fi 89: } 90: 91: 92: # If a host responds to pings, it is considered up. 93: function checkconnection() { 94: 95: unset UP 96: 97: # ping host 98: PING_ANSWER="$($PING -R -B 2 -r $RETRIES -p 50 "$HOST" 2>&1)" 99: PING_ERRNUM=$? 100: if [ $PING_ERRNUM -gt 2 ]; then 101: echo "Error: Got error $PING_ERRNUM from fping $(head -1 "$PING_ANSWER"). Disregarding $HOST." >&2 102: continue; 103: fi 104: 105: # parse answer 106: if [ $(echo "$PING_ANSWER" | grep -c "$ALIVE_ANSWER") -gt 0 ]; then 107: UP=true 108: fi 109: TIME_STAMP="$(date +"%a %e.%m.%y %H:%M:%S")" 110: if [ -n "$UP" ]; then 111: if [ ! "$DEBUG" = "low" ] && [ -n "$DEBUG" ]; then 112: shout "The connection to $HOST is up." 113: fi 114: else 115: if [ -n "$DEBUG" ]; then 116: shout "The connection to $HOST is down." 117: fi 118: fi 119: 120: # remember a hosts state 121: rememberstate 122: } 123: 124: 125: # Rember recent state of host, trigger alert 126: function rememberstate() { 127: # Put a simple time stamp in a tmp file, when host is detected as down for the first time 128: if [ -z "$UP" ]; then 129: if [ ! -e "$TMPDIR/.canardien-$PINGHOST-$HOST" ]; then 130: # set time stamp 131: echo -ne "$TIME_STAMP" > "$TMPDIR/.canardien-$PINGHOST-$HOST" 132: # send alert 133: sendalert 134: elif [ -n "$DEBUG" ]; then 135: shout "Down since `cat $TMPDIR/.canardien-$PINGHOST-$HOST`." 136: fi 137: else 138: # Host is up 139: if [ -e "$TMPDIR/.canardien-$PINGHOST-$HOST" ]; then 140: if [ -n "$DEBUG" ]; then 141: shout "$HOST is up again." 142: fi 143: # send alert 144: ALERT_SUBJECT='"The host $HOST answers again!"' 145: ALERT_TEXT='"[$(date +"%a %e.%m.%y %H:%M:%S")]\n\n\n$HOST is up again.\n--------------------------------------\n(downtime began $(cat "$TMPDIR/.canardien-$PINGHOST-$HOST"))\n\n\n Kind regards, $PINGHOST."' 146: sendalert 147: # delete tmp file 148: rm "$TMPDIR/.canardien-$PINGHOST-$HOST" 149: else 150: # still running 151: if [ ! "$DEBUG" = "low" ] && [ -n "$DEBUG" ]; then 152: shout "$HOST is up." 153: fi 154: fi 155: fi 156: } 157: 158: 159: # Send an email alert 160: function sendalert() { 161: # send mail 162: if [ ! -z "$ALERT_TO" ]; then 163: eval alert_subject=\$$ALERT_SUBJECT 164: eval MSG=\$$ALERT_TEXT 165: echo -e -n "$MSG" | mail -s "$alert_subject" "$ALERT_TO" 166: if [ -n "$DEBUG" ]; then 167: shout "Alert sent to $ALERT_TO." 168: fi 169: fi 170: } 171: 172: 173: # To avoid any output in case of a silent operation, 174: # shout instead of echo. 175: function shout() { 176: if [ -z "$SILENT" ]; then 177: echo -e "$1" 178: fi 179: } 180: 181: 182: # -----------main----------- 183: 184: init "$*" 185: 186: # call an instance for each host, when multiple hosts are given 187: if [[ "$HOSTS" =~ " " ]]; then 188: while read -r HOST; do 189: "$0" "$HOST" & 190: done <<< "$(echo "$HOSTS" | tr ' ' '\n')" 191: exit 0 192: else 193: # process a single host 194: HOST="$HOSTS" 195: checkconnection 196: fi 197: 198: exit 0 199: 200: # todo: Switches ( -q, --debug, ...) 201: # grace period: do not send a mail if down time shorter than X minutes 202: # repeat down info after x hours 203: # test port instead of ping or in addition 204: