check_newest_file_age

All is good here! Thanks for your work

---

Hola,

He realizado algunos ajustes para que el plugin funcione en Solaris 10. Ya lo he validado y ya se encuentra en producción.

Nota: disculpen el no escribir en ingles, ya que soy de habla hispana y prefiero no complicarme con las traducciones. ;-)

A continuación copio todo el codigo del script:

PROGPATH=`dirname $0`
REVISION=`echo '$Revision: 1 $' | sed -e 's/[^0-9.]//g'`

. $PROGPATH/utils.sh

print_usage() {
echo "
Usage: check_newest_file_age --dirs | -d [-w ] [-c ] [-W] [-C] [-t ] [-V] [--check-dirs] [--base-dir ]
Usage: check_newest_file_age --help | -h

Description:

This plugin pulls the most recently created file in each specified directory,
and checks it's created time against the current time. If the maximum age of
the file is exceeded, a warning/critical message is returned as appropriate.

This is useful for examining backup directories for freshness.

Tested to work on Linux/FreeBSD/OS X.

The following arguments are accepted:

--dirs | -d A space separated list of directories to examine. Each
directory will be checked for the newest created file in that
directory.

-w (Optional) Generate a warning message if the last created
file is older than this value. Defaults to 26 hours.

-c (Optional) Generate a critical message if the last created
file is older than this value. Defaults to 52 hours.

-W (Optional) If set, a warning message will be returned if the
specified directory doesn't exist, or there are no checkable
files in the specified directory.

-C (Optional) If set, a critical message will be returned if the
specified directory doesn't exist, or there are no checkable
files in the specified directory.

-t (Optional) The time unit used for the -w and -c values. Must
be one of: seconds, minutes, hours, days. Defaults to hours.

-V (Optional) Output verbose information about all checked
directories. Default is only to print verbose information
for directories with non-OK states.

--check-dirs (Optional) If set, directories inside the specified directory
will also be checked for their creation time. Note that this
check is not recursive. Without this option, only real files
inside the specified directory will be checked.

--base-dir (Optional) If set, this path will be prepended to all
checked directories.

--help | -h Print this help and exit.

Examples:

Generate a warning if the newest file in /backups is more than 26 hours old,
and a critical if it's more than 52 hours old...

check_newest_file_age -d "/backups"

Generate a warning if the newest file in /backups/bill or /backups/dave is more
than one week old, or a critical if it's more than two weeks old...

check_newest_file_age -d "/backups/bill /backups/dave" -w 7 -c 14 -t days

Caveats:

Although multiple directories can be specified, only one set of
warning/critical times can be supplied.

Linux doesn't seem to have an easy way to check file/directory creation time,
so file/directory last modification time is used instead.
"
}

print_help() {
print_usage
echo "Newest file in a directory plugin for Nagios."
echo ""
}

# Sets the exit status for the plugin. This is done in such a way that the
# status can only go in one direction: OK -> WARNING -> CRITICAL.
set_exit_status() {
new_status=$1
# Nothing needs to be done if the state is already critical, so exclude
# that case.
case $exitstatus
in
$STATE_WARNING)
# Only upgrade from warning to critical.
if [ "$new_status" = "$STATE_CRITICAL" ]; then
exitstatus=$new_status;
fi
;;
$STATE_OK)
# Always update state if current state is OK.
exitstatus=$new_status;
;;
esac
}

# Make sure the correct number of command line
# arguments have been supplied
if [ $# -lt 1 ]; then
print_usage
exit $STATE_UNKNOWN
fi

# Defaults.
exitstatus=$STATE_OK
warning=26
critical=52
time_unit=hours
verbose=
on_empty=$STATE_OK
check_dirs=
base_dir=

# Grab the command line arguments.
while test -n "$1"; do
case "$1" in
--help)
print_help
exit $STATE_OK
;;
-h)
print_help
exit $STATE_OK
;;
--dirs)
dirs=$2
shift
;;
-d)
dirs=$2
shift
;;
-w)
warning=$2
shift
;;
-c)
critical=$2
shift
;;
-W)
on_empty=$STATE_WARNING
;;
-C)
on_empty=$STATE_CRITICAL
;;
-t)
time_unit=$2
shift
;;
-V)
verbose=1
;;
--check-dirs)
check_dirs=1
;;
--base-dir)
base_dir=$2
shift
;;
-x)
exitstatus=$2
shift
;;
--exitstatus)
exitstatus=$2
shift
;;
*)
echo "Unknown argument: $1"
print_usage
exit $STATE_UNKNOWN
;;
esac
shift
done

if [ ! "$dirs" ]; then
echo "No directories provided."
exit $STATE_UNKNOWN
fi

eval=`expr $warning + 0`
if [ -z $eval ]; then
echo "Warning value must be a number."
exit $STATE_UNKNOWN
fi

eval=`expr $critical + 0`
if [ -z $eval ]; then
echo "Critical value must be a number."
exit $STATE_UNKNOWN
fi

if [ ! `echo "$time_unit" | egrep "seconds|minutes|hours|days"` ]; then
echo "Time unit must be one of: seconds, minutes, hours, days."
exit $STATE_UNKNOWN
fi

if [ "$warning" -ge "$critical" ]; then
echo "Critical time must be greater than warning time."
exit $STATE_UNKNOWN
fi

case $time_unit
in
days)
multiplier=86400;
abbreviation="days";
;;
hours)
multiplier=3600;
abbreviation="hrs";
;;
minutes)
multiplier=60;
abbreviation="mins";
;;
*)
multiplier=1
abbreviation="secs";
;;
esac

# Starting values.
DIR_COUNT=0
OK_FILE_COUNT=0
OUTPUT=
CURRENT_TIME=`perl -e 'print time."n"'`
OS_DISTRO=`uname -s`

# Loop through each provided directory.
for dir in $dirs
do
check_file=
let "DIR_COUNT += 1" > /dev/null

# Check if dir exists.
full_path=${base_dir}${dir}
if [ -d "$full_path" ]; then
file_list=`ls -t $full_path`
# Cycle through files, looking for a checkable file.
for next_file in $file_list
do
next_filepath=$full_path/$next_file
if [ "$check_dirs" ]; then
# Check if it's a file or directory.
if [ -f "$next_filepath" ] || [ -d "$next_filepath" ]; then
check_file=1
fi
else
# Check if it's a file.
if [ -f "$next_filepath" ]; then
check_file=1
fi
fi
if [ "$check_file" ]; then
# stat doesn't work the same on Linux and FreeBSD/Darwin, so
# make adjustments here.
if [ "$OS_DISTRO" = "Linux" ]; then
st_ctime=`stat --printf=%Y ${next_filepath}`
else
#eval $(stat -s ${next_filepath})
st_ctime=`perl -e 'print scalar((stat $ARGV[0])[9])' ${next_filepath}`
fi

let "FILE_AGE = $CURRENT_TIME - $st_ctime" > /dev/null
let "FILE_AGE_UNITS = $FILE_AGE / $multiplier" > /dev/null
let "MAX_WARN_AGE = $warning * $multiplier" > /dev/null
let "MAX_CRIT_AGE = $critical * $multiplier" > /dev/null
if [ $FILE_AGE -gt $MAX_CRIT_AGE ]; then
OUTPUT="$OUTPUT ${dir}: ${FILE_AGE_UNITS}${abbreviation}"
set_exit_status $STATE_CRITICAL
elif [ $FILE_AGE -gt $MAX_WARN_AGE ]; then
OUTPUT="$OUTPUT ${dir}: ${FILE_AGE_UNITS}${abbreviation}"
set_exit_status $STATE_WARNING
else
let "OK_FILE_COUNT += 1" > /dev/null
if [ "$verbose" ]; then
OUTPUT="$OUTPUT ${dir}: ${FILE_AGE_UNITS}${abbreviation}"
fi
fi
break
fi
done
# Check here to see if any files got tested in the directory.
if [ ! "$check_file" ]; then
set_exit_status $on_empty
OUTPUT="$OUTPUT ${dir}: No files"
# If empty is an OK state, then increment the ok file count.
if [ "$on_empty" = "$STATE_OK" ]; then
let "OK_FILE_COUNT += 1" > /dev/null
fi
fi
else
set_exit_status $on_empty
OUTPUT="$OUTPUT ${dir}: Does not exist"
fi
done

case $exitstatus
in
$STATE_CRITICAL)
exit_message="CRITICAL";
;;
$STATE_WARNING)
exit_message="WARNING";
;;
$STATE_OK)
exit_message="OK";
;;
*)
exitstatus=$STATE_UNKNOWN;
exit_message="UNKNOWN";
;;
esac

exit_message="${exit_message}: ${OK_FILE_COUNT}/${DIR_COUNT}"

if [ "$OUTPUT" ]; then
exit_message="${exit_message} --${OUTPUT}"
fi

echo "$exit_message"
exit $exitstatus

---

If you want to do it recursive, use find to locate the indicated directory most recent file (on the tree).
Then, take the file, and break the loop, indicate that the next file (the most recent unique file) is the find result file.

Here is an example code, of modificated lines, excluding *.ini, *.tm? and *.db files:

# Loop through each provided directory.
for dir in $dirs
do
check_file=
let "DIR_COUNT += 1" > /dev/null

# Check if dir exists.
full_path=${base_dir}${dir}

#Agregado por Cristian Barreto
full_path_archivo=`find $full_path -not -name "*.tm?" -not -name "*.ini" -not -name "*.db" $
indiceBarra=`expr match "$full_path_archivo" ".*/"`
full_path=`echo $full_path_archivo | cut -c 1-$indiceBarra`
#FIN del agregado
if [ -d "$full_path" ]; then
file_list=$full_path_archivo
# Cycle through files, looking for a checkable file.
for next_file in "$file_list"
do next_filepath=$full_path_archivo
if [ "$check_dirs" ]; then
# Check if it's a file or directory.
if [ -f "$next_filepath" ] || [ -d "$next_filepath" ]; then
check_file=1
fi

.
.
.
.
break #at the end of loop
done

--------------------------------

done!, remember that $next_filepath must have "$next_filepath" and "${next_filepath}" to support file names with blank spaces like Windows file names.

Sorry for my bad english.

---

I give this one a big 5 stars. Code quality wise, it is one of the better plugins I have ran across. The author did a really nice job.

We have this checking a number of directory trees, to verify that the backup files are getting dropped into there on schedule - as soon as the newest file is 24+ hours old, it implies that the job didn't complete successfully and an alert is fired. Works great.

---