Hello I created a script to kill processes ordered by age however everytime the PIDs are changing... how can I solve this
here is my script
#!/bin/bash
#Argument = -c check -k kill -l list
usage()
{
cat << EOF
usage: $0 options
This script kills all the processes running and leaves the last one sorted by age running.
OPTIONS:
-c checks how many proccess are runnig it needs string argument
-k Kill all the processes and leaves just the last sorted by age running
-l Show the list of procesess to be killed.
EOF
}
CHECK=
KILL=
LIST=
while getopts "hc:k:l:" OPTION
do
case $OPTION in
h)
usage
exit 1
;;
c)
CHECK=$OPTARG
ps -ef | grep -i $CHECK | wc -l
;;
k)
KILL=$OPTARG
T2=$(ps -ef | grep -i "$KILL" | awk '{print $3,$5}' | sort -r +1 | sed 1d |awk '{print $1}')
for f in $T2; do
echo "killing $f"
kill $f
done
;;
l)
LIST=$OPTARG
T2=$(ps -ef | grep -i "$LIST" | awk '{print $3,$5}' | sort -r +1 | sed 1d |awk '{print $1}')
for f in $T2; do
echo "PID $f"
done
;;
?)
usage
exit
;;
esac
done
if [[ -z KILL ]] || [[ -z LIST ]] || [[ -z CHECK ]]
then
usage
exit 1
fi
and also I don't understand why when I call the script with no arguments the help doesn't show up
The PIDs will change if another program is restarting it when it's killed. This is actually pretty common with daemons.
usage is never called because you're checking whether the strings KILL etc. are empty, not the variables. Just add a dollar sign in front of them.
Related
I'm trying to detect if 1 user has 2 terminal sessions open.
How can this be achieved?
What I tried is the following:
linuxName="$(id -u -n)"
for user in "$#"; do
if echo "$who" | grep -q "$linuxName"; then
echo "[WARNING] You're already logged in";
closeSessionIfNotRoot
fi;
done;
sort count and filter these that have more then 2.
id -u -n | sort | uniq -c | awk '$1 > 2'
#!/usr/bin/bash
linuxName="$(id -un)"
for user in "$#"; do
process_counter=$(pgrep -u "$user" bash 2>&1 | wc -l)
if [[ "$user" == $(who | grep -o "$linuxName") ]]; then
echo "[WARNING] You're already logged in $user"
if [[ "$process_counter" -gt 1 ]]; then
((process_counter--))
pgrep -u "$user" bash | tail -"$process_counter" |
xargs -I % kill -9 %
echo "closed other terminals: $user"
fi
fi
done
script +arguments[in this case, users_names], example, ./script alex zara sophia
Explain answer
Count terminals open by user and save to variable process_counter
process_counter=$(pgrep -u "$user" bash 2>&1 | wc -l)
2>1 REDIRECTS output error to NO SHOW. this in case you wrote a user_name that doesn't exist in the system.
pgrep -u $user bash | wc -l returns number of terminals open. u $user is for returning the process from that user, so this uniq.
pgrep shows the process you are searching for,e.g., pgreg bash returns a lists of process IDs
returns:
363570
369836
370113
this checks if the $user --> ./file alex zara is the same as in terminal $linuxName
[[ "$user" == $(who | grep -o "$linuxName") ]]
grep -o only returns the username alex instead of alex :1 2021-07....
[[ ]] is to make comparisons in bash
xargs is for pasing previous argument to a loop
tail -"$process_counter" | xargs -I % kill -9 %
tail -$process_counter is for showing the last terminals open
363570
369836 #return only this
370113 #return only this
remember that we REST minus 1 to process_counter because we want to leave 1 terminal open, this is done with ((process_counter--)).
So 3 terminals open - 1 terminal = 2 terminals to close.
now we loop with XARGS
363570
369836 #CLOSE THIS kill -9 369836
370113 #CLOSE THIS kill -9 370113
-I arguments tells xargs to assign a variable, represented with symbol %. but this symbol can be anything,like #,$, and then what we want to do is kill each process, so that's why kill -9 %
Check limits.conf which is configuration file for the pam_limits module.
In that file, you can define maxlogins as the maximum number of logins for a user (this limit does not apply to user with uid=0)
See https://www.man7.org/linux/man-pages/man5/limits.conf.5.html.
How can I make the following commands exit immediately after the first line is matched? I understand that SIGPIPE is not sent to cat until it tries to write next time (tail bug report), but I don't understand how to solve this issue.
cat <( echo -ne "asdf1\nzxcv1\nasdf2\n"; sleep 5; echo -ne "zxcv2\nasdf3\n" ) | grep --line-buffered zxcv | head --lines=1
cat <( echo -ne "asdf1\nzxcv1\nasdf2\n"; sleep 5; echo -ne "zxcv2\nasdf3\n" ) | grep --max-count=1 zxcv
NB: I actually had tail --follow before the pipesign, but replaced it with catand sleep to simplify testing. The shell in question is GNU bash 4.4.12(1)-release, and I'm running MINGW that came with Git-for-Windows 2.12.2.2.
CLARIFICATION: I have a jboss server which is started in a docker container and which outputs couple thousand lines of text within three minutes to a log file. My goal is to watch this file until a status line is printed, analyze line contents and return it to a human or Jenkins user. Of course, I can grep whole file and sleep for a second in a loop, but I'd rather avoid this if at all possible. Furthermore, this looping would interfere with my usage of timeout routine to limit maximum execution time. So, is it possible to listen for a pipe until a certain line appears and stop as soon as that happens?
Related question: Why does bash ignore SIGINT if its currently running child handles it?
Interesting question! I've verified that head dies after printing the first line (removed background job noise):
$ (printf '%s\n' a b a; sleep 5; printf '%s\n' a) | grep --line-buffered a | head --lines=1 & sleep 1; pstree $$
a
bash─┬─bash───sleep
├─grep
└─pstree
At first glance, it appears head doesn't send SIGPIPE, but I get conflicting information from running strace grep:
$ (printf '%s\n' a b a; sleep 10; printf '%s\n' a) | strace grep --line-buffered a | head --lines=1
…
--- SIGPIPE {si_signo=SIGPIPE, si_code=SI_USER, si_pid=21950, si_uid=1000} ---
+++ killed by SIGPIPE +++
… and killing grep:
$ (printf '%s\n' a b a; sleep 10; printf '%s\n' a) | grep --line-buffered a | head --lines=1 & sleep 1; kill -PIPE $(pgrep grep); sleep 5; pstree $$
a
bash─┬─bash───sleep
└─pstree
Killing grep and then sleep fixes the issue:
$ (printf '%s\n' a b a; sleep 10; printf '%s\n' a) | grep --line-buffered a | head --lines=1 & sleep 1; kill -PIPE $(pgrep grep); sleep 1; kill -PIPE $(pgrep sleep); sleep 5; pstree $$
a
bash───pstree
Conclusion: WTF?
I've ended up doing following to be able to break following log both on a matching line and after a timeout.
#!/bin/sh
TOP_PID=$$
container_id="$1"
LOG_PATH=/opt/jboss/jboss-eap-6.2/standalone/log/server.log
await_startup () {
status=$(check_status)
follow_log --timeout $timeout &
local bgjob_pid; local bgjob_status;
bgjob_pid=$(jobs -p)
test -n "$bgjob_pid" || die "Could not start background job to follow log."
bgjob_status=true
while [ "$status" = "not started" ] && $bgjob_status; do
sleep 1s
status=$(check_status)
if kill -0 $bgjob_pid 2>/dev/null; then
bgjob_status=true
else
bgjob_status=false
fi
done
kill -KILL $bgjob_pid 2>/dev/null
}
follow_log () {
# argument parsing skipped...
docker exec $container_id timeout $timeout tail --follow=name ---disable-inotify --max-unchanged-stats=2 /$LOG_PATH
}
check_status () {
local line;
line=$(docker exec $container_id grep --extended-regexp --only-matching 'JBoss EAP .+ started.+in' /$LOG_PATH | tail --lines=1)
if [ -z "$line" ]; then
printf "not started"
elif printf "%s" "$line" | grep --quiet "with errors"; then
printf "started and unhealthy"
else
printf "healthy"
fi
}
die () {
test -n "$1" && printf "%s\n" "$1"
kill -s TERM $TOP_PID
return 1
} 1>&2
I've written the bash script (searchuser) which should display all the users who are executing a specific program or a script (at least a bash script). But when searching for scripts fails because the command the SO is executing is something like bash scriptname.
This script acts parsing the ps command output, it search for all the occurrences of the specified program name, extracts the user and the program name, verifies if the program name is that we're searching for and if it's it displays the relevant information (in this case the user name and the program name, might be better to output also the PID, but that is quite simple). The verification is accomplished to reject all lines containing program names which contain the name of the program but they're not the program we are searching for; if we're searching gedit we don't desire to find sgedit or gedits.
Other issues I've are:
I would like to avoid the use of a tmp file.
I would like to be not tied to GNU extensions.
The script has to be executed as:
root# searchuser programname <invio>
The script searchuser is the following:
#!/bin/bash
i=0
search=$1
tmp=`mktemp`
ps -aux | tr -s ' ' | grep "$search" > $tmp
while read fileline
do
user=`echo "$fileline" | cut -f1 -d' '`
prg=`echo "$fileline" | cut -f11 -d' '`
prg=`basename "$prg"`
if [ "$prg" = "$search" ]; then
echo "$user - $prg"
i=`expr $i + 1`
fi
done < $tmp
if [ $i = 0 ]; then
echo "No users are executing $search"
fi
rm $tmp
exit $i
Have you suggestion about to solve these issues?
One approach might looks like such:
IFS=$'\n' read -r -d '' -a pids < <(pgrep -x -- "$1"; printf '\0')
if (( ! ${#pids[#]} )); then
echo "No users are executing $1"
fi
for pid in "${pids[#]}"; do
# build a more accurate command line than the one ps emits
args=( )
while IFS= read -r -d '' arg; do
args+=( "$arg" )
done </proc/"$pid"/cmdline
(( ${#args[#]} )) || continue # exited while we were running
printf -v cmdline_str '%q ' "${args[#]}"
user=$(stat --format=%U /proc/"$pid") || continue # exited while we were running
printf '%q - %s\n' "$user" "${cmdline_str% }"
done
Unlike the output from ps, which doesn't distinguish between ./command "some argument" and ./command "some" "argument", this will emit output which correctly shows the arguments run by each user, with quoting which will re-run the given command correctly.
What about:
ps -e -o user,comm | egrep "^[^ ]+ +$1$" | cut -d' ' -f1 | sort -u
* Addendum *
This statement:
ps -e -o user,pid,comm | egrep "^\s*\S+\s+\S+\s*$1$" | while read a b; do echo $a; done | sort | uniq -c
or this one:
ps -e -o user,pid,comm | egrep "^\s*\S+\s+\S+\s*sleep$" | xargs -L1 echo | cut -d ' ' -f1 | sort | uniq -c
shows the number of process instances by user.
Welcome, I have a short script to kill processes which works longer than specified time for UIDs bigger than. How to exclude for example mc command from killing?
#!/bin/bash
#
#Put the minimum(!) UID to kill processes
UID_KILL=500
#Put the time in seconds which the process is allowed to run below
KILL_TIME=300
KILL_LIST=`{
ps -eo uid,pid,lstart | tail -n+2 |
while read PROC_UID PROC_PID PROC_LSTART; do
SECONDS=$[$(date +%s) - $(date -d"$PROC_LSTART" +%s)]
if [ $PROC_UID -ge $UID_KILL -a $SECONDS -gt $KILL_TIME ]; then
echo -n "$PROC_PID "
fi
done
}`
#KILLING LOOP
while sleep 1
do
if [[ -n $KILL_LIST ]]
then
kill $KILL_LIST
fi
done
change inner command like this :
ps -eo comm,uid,pid,lstart | tail -n+2 | grep -v '^your_command' | ...
this will exclude 'your_command' from the list.
see STANDARD FORMAT SPECIFIERS in man ps for more about ps -o.
First off, I'm new to this. I have some experience with windows scripting and apple script but not much with bash. What I'm trying to do is grab the PID and %CPU of a specific process. then compare the %CPU against a set number, and if it's higher, kill the process. I feel like I'm close, but now I'm getting the following error:
[[: 0.0: syntax error: invalid arithmetic operator (error token is ".0")
what am I doing wrong? here's my code so far:
#!/bin/bash
declare -i app_pid
declare -i app_cpu
declare -i cpu_limit
app_name="top"
cpu_limit="50"
app_pid=`ps aux | grep $app_name | grep -v grep | awk {'print $2'}`
app_cpu=`ps aux | grep $app_name | grep -v grep | awk {'print $3'}`
if [[ ! $app_cpu -gt $cpu_limit ]]; then
echo "crap"
else
echo "we're good"
fi
Obviously I'm going to replace the echos in the if/then statement but it's acting as if the statement is true regardless of what the cpu load actually is (I tested this by changing the -gt to -lt and it still echoed "crap"
Thank you for all the help. Oh, and this is on a OS X 10.7 if that is important.
I recommend taking a look at the facilities of ps to avoid multiple horrible things you do.
On my system (ps from procps on linux, GNU awk) I would do this:
ps -C "$app-name" -o pid=,pcpu= |
awk --assign maxcpu="$cpu_limit" '$2>maxcpu {print "crappy pid",$1}'
The problem is that bash can't handle decimals. You can just multiply them by 100 and work with plain integers instead:
#!/bin/bash
declare -i app_pid
declare -i app_cpu
declare -i cpu_limit
app_name="top"
cpu_limit="5000"
app_pid=`ps aux | grep $app_name | grep -v grep | awk {'print $2'}`
app_cpu=`ps aux | grep $app_name | grep -v grep | awk {'print $3*100'}`
if [[ $app_cpu -gt $cpu_limit ]]; then
echo "crap"
else
echo "we're good"
fi
Keep in mind that CPU percentage is a suboptimal measurement of application health. If you have two processes running infinite loops on a single core system, no other application of the same priority will ever go over 33%, even if they're trashing around.
#!/bin/sh
PROCESS="java"
PID=`pgrep $PROCESS | tail -n 1`
CPU=`top -b -p $PID -n 1 | tail -n 1 | awk '{print $9}'`
echo $CPU
I came up with this, using top and bc.
Use it by passing in ex: ./script apache2 50 # max 50%
If there are many PIDs matching your program argument, only one will be calculated, based on how top lists them. I could have extended the script by catching them all and avergaing the percentage or something, but this will have to do.
You can also pass in a number, ./script.sh 12345 50, which will force it to use an exact PID.
#!/bin/bash
# 1: ['command\ name' or PID number(,s)] 2: MAX_CPU_PERCENT
[[ $# -ne 2 ]] && exit 1
PID_NAMES=$1
# get all PIDS as nn,nn,nn
if [[ ! "$PID_NAMES" =~ ^[0-9,]+$ ]] ; then
PIDS=$(pgrep -d ',' -x $PID_NAMES)
else
PIDS=$PID_NAMES
fi
# echo "$PIDS $MAX_CPU"
MAX_CPU="$2"
MAX_CPU="$(echo "($MAX_CPU+0.5)/1" | bc)"
LOOP=1
while [[ $LOOP -eq 1 ]] ; do
sleep 0.3s
# Depending on your 'top' version and OS you might have
# to change head and tail line-numbers
LINE="$(top -b -d 0 -n 1 -p $PIDS | head -n 8 \
| tail -n 1 | sed -r 's/[ ]+/,/g' | \
sed -r 's/^\,|\,$//')"
# If multiple processes in $PIDS, $LINE will only match\
# the most active process
CURR_PID=$(echo "$LINE" | cut -d ',' -f 1)
# calculate cpu limits
CURR_CPU_FLOAT=$(echo "$LINE"| cut -d ',' -f 9)
CURR_CPU=$(echo "($CURR_CPU_FLOAT+0.5)/1" | bc)
echo "PID $CURR_PID: $CURR_CPU""%"
if [[ $CURR_CPU -ge $MAX_CPU ]] ; then
echo "PID $CURR_PID ($PID_NAMES) went over $MAX_CPU""%"
echo "[[ $CURR_CPU""% -ge $MAX_CPU""% ]]"
LOOP=0
break
fi
done
echo "Stopped"
Erik, I used a modified version of your code to create a new script that does something similar. Hope you don't mind it.
A bash script to get the CPU usage by process
usage:
nohup ./check_proc bwengine 70 &
bwegnine is the process name we want to monitor 70 is to log only when the process is using over 70% of the CPU.
Check the logs at: /var/log/check_procs.log
The output should be like:
DATE | TOTAL CPU | CPU USAGE | Process details
Example:
03/12/14 17:11 |20.99|98| ProdPROXY-ProdProxyPA.tra
03/12/14 17:11 |20.99|100| ProdPROXY-ProdProxyPA.tra
Link to the full blog:
http://felipeferreira.net/?p=1453
It is also useful to have app_user information available to test whether the current user has the rights to kill/modify the running process. This information can be obtained along with the needed app_pid and app_cpu by using read eliminating the need for awk or any other 3rd party parser:
read app_user app_pid tmp_cpu stuff <<< \
$( ps aux | grep "$app_name" | grep -v "grep\|defunct\|${0##*/}" )
You can then get your app_cpu * 100 with:
app_cpu=$((${tmp_cpu%.*} * 100))
Note: Including defunct and ${0##*/} in grep -v prevents against multiple processes matching $app_name.
I use top to check some details. It provides a few more details like CPU time.
On Linux this would be:
top -b -n 1 | grep $app_name
On Mac, with its BSD version of top:
top -l 1 | grep $app_name