I have user1 run a script called logginexpert.sh while has this line of code
sleep 888
I then login to another putty session with another user2 and try to grep for the logginexpert.sh process using
ps -ef | grep ex
SunOS mymac 5.11 11.2 sun4u sparc SUNW,SPARC-Enterprise
But, i dont get any results when i was expecting it to show the process logginexpert.sh that was started by user1.
Can you please help me fix this problem ?
Does it show up in the long list, when not grep ping?
Not sure what long list is but i grep for the entire script name and it still does not showup in the results.
ps -ef | grep logginexpert.sh
No results Found.
I also tried
ps -ef >/tmp/tmp.tmp
grep ex /tmp/tmp.tmp
No results Found.
Can you even SEE the process using ps -ef
Do not use more or grep. Why? If the path to the file is part of the ps display for
logginexpert.sh
then it may include the entire path - which can exceed what ps will show - too many characters.
PS: on Solaris pgrep may be your better choice for this kind of thing.
location of logginexpert.sh is /tmp/logginexpert.sh
ps -ef | pgrep /tmp/testbkpme.sh
No results Found.
ps -ef | pgrep bk
No results Found.
ps -ef also does not show /tmp/logginexpert.sh but I'm able to grep the process by the pid but not with the script name.
Putty Session 1:
/tmp/logginexpert.sh &
[1] 4627
Putty Session 2:
You have new mail in /var/mail/user2
ps -ef | grep 4627
user1 4627 10477 0 13:54:09 pts/11 0:00 sleep 888
ps -ef | grep ex
user2 5650 15730 0 13:54:27 pts/16 No results Found.
So definitely the process is running.
---------- Post updated at 03:12 PM ---------- Previous update was at 02:36 PM ----------
Also ps -ef does not show the name of the script but i see this sleep 888 in the results.
Can you please help sort this out ? i want to grep by name of the script not by its contents.
Display the processes of user1
ps -fu user1
If "sleep" is spawned by the script then look for its parent process (given in column 3).
Maybe your grep is forged?
/bin/ps -fu user1 | /bin/grep ex
grep can match its own argument in the ps list. pgrep avoids that.
pgrep -u user1 ex
The issue is not to do with different users or different putty session. Becoz i Cannot find i.e grep for the logginexpert.sh script with the same user in the same putty session.
Also,
/bin/ps -fu | /bin/grep ex
/bin/ps: option requires an argument � u
pgrep -u user1 ex
No results Found.
/bin/ps fu | /bin/grep ex
/bin/ps: illegal option -- f
/bin/ps u | /bin/grep bk
Result not Found.
Let me know what are we missing ?
How do you expect grep to find it, then? Please show the output of /bin/ps -ef for ALL the relevant processes.
Here is the Output
UID PID PPID C STIME TTY TIME CMD
netcfg 5865 4775 0 Nov 20 ? 0:36 /lib/inet/netcfgd
root 4775 4775 0 Nov 20 ? 0:00 zsched
daemon 5887 4775 0 Nov 20 ? 0:17 /lib/crypto/kcfd
root 5997 4775 0 Nov 20 ? 1:17 /lib/inet/in.mpathd
user2 7364 5339 0 04:01:03 ? 0:00 ksh -x pls_med_02001.sh
root 12094 11832 0 10:49:27 pts/5 0:00 sudo -u user2 -i -H
user2 13301 13290 0 Nov 28 pts/10 0:00 zsh
root 11047 4775 0 Nov 20 ? 66:02 ./Agent.bin -a
root 12877 12720 0 Nov 28 pts/9 0:00 sudo -u user2 -s -H
root 12239 12094 0 10:49:30 pts/6 0:00 sudo -u user2 -i -H
user2 16470 29038 0 04:04:28 ? 0:00 sleep 60
user1 26410 26381 0 12:04:26 pts/17 0:00 bash
user2 15504 15503 0 04:03:44 pts/16 0:00 sleep 888
user1 26381 26380 0 12:04:24 pts/17 0:00 -ksh
user1 28576 28575 0 12:05:37 pts/8 0:00 -ksh
u1064978 12720 12433 0 Nov 28 pts/9 0:00 bash
root 15626 15477 0 12:11:39 pts/16 0:00 sudo -u user2 -s -H
user2 15730 15627 0 12:11:41 pts/16 0:00 bash
user2 16716 16712 0 04:04:40 ? 0:00 tee -a sendemail.603486944.154.log
root 10476 10328 0 12:21:43 pts/11 0:00 sudo -u user2 -s -H
root 2512 2511 0 00:33:22 ? 0:01 cdcwatch 2511 -F -M
user2 12241 12239 0 10:49:30 pts/6 0:00 -ksh
user2 25389 25384 0 03:40:07 ? 0:00 ksh -x eload.sh 431 LOAD REAL 58731
root 10328 26410 0 12:21:39 pts/17 0:00 sudo -u user2 -s -H
rgopi 9324 4775 0 - ? 0:00 <defunct>
root 13289 12877 0 Nov 28 pts/10 0:00 sudo -u user2 -s -H
root 15477 781 0 12:11:36 pts/8 0:00 sudo -u user2 -s -H
user2 16256 28862 0 04:04:23 ? 0:00 sleep 60
user2 15503 15730 0 04:03:44 pts/16 0:00 bash
user2 12309 12241 0 10:49:30 pts/6 0:00 zsh
user2 15949 12309 0 10:50:45 pts/6 0:00 cat
user1 781 28576 0 12:06:36 pts/8 0:00 bash
You have new mail in /var/mail/user1
The one in BOLD is my process which i m not able to grep (find).
Which "ex" string do you want to match in that bolded line?
That process' parent process (ID 15503) is a bash shell - NO logginexpert.sh ! So I have to challenge your first statement in post#1.
k ... looks like i have to explain this again.
The script name is logginexpert.sh which has the sleep 888 line in it.
So, when i run logginexpert.sh in the background i see the sleep 888 which is shown in the output of my /bin/ps -ef command.
So, it is not listing the process with the name of the script i.e logginexpert.sh which is what i want !! ...instead it is listing the sleep 888 command that is in the script.
I have highlighted this before as well.
And I'd challenge that statement. A script would usually wait for its children (and then show up in the ps listing). I don't even believe you put sleep 888 into background with nohup or disown , as then, it's parent would become 1 ( init ).
Post the relevant part of logginexpert.sh .
more logginexpert.sh
echo wowww
sleep 888
echo wow
You are probably hitting the problem that you script is only showing as a bash process like me:-
Session 1:-
$ cat /tmp/robinsleep.sh
sleep 888
$ /tmp/robinsleep.sh
Session 2:
$ ps -ef|grep sleep
rbatte1 21561 21560 0 11:47 pts/9 00:00:00 sleep 888
rbatte1 21563 21537 0 11:47 pts/10 00:00:00 grep sleep
$ ps -fp 21560
UID PID PPID C STIME TTY TIME CMD
rbatte1 21560 21516 0 11:47 pts/9 00:00:00 bash
$
Robin
k... whats the solution ?
Well, that depends on what the overall process is. What are you trying to work out?
For instance:-
if we can understand the context, then we may be able to suggest a better solution that will be flexible and let you evolve and support the code for the future.
Robin
Interesting difference between script with "shebang" and one without:
cat shsleep1
#! /bin/bash
sleep 888
./shsleep1
ps -efH
.
.
.
3272 3219 0 80 0 - 5698 wait 10:38 pts/3 00:00:00 /bin/bash
4423 3272 0 80 0 - 3518 wait 13:12 pts/3 00:00:00 /bin/bash ./shsleep1
4424 4423 0 80 0 - 2205 hrtime 13:12 pts/3 00:00:00 sleep 888
cat shsleep2
sleep 888
./shsleep2
ps -efH
.
.
.
3272 3219 0 80 0 - 5698 wait 10:38 pts/3 00:00:00 /bin/bash
4588 3272 0 80 0 - 5701 wait 13:15 pts/3 00:00:00 /bin/bash
4589 4588 0 80 0 - 2205 hrtime 13:15 pts/3 00:00:00 sleep 888
I didn't realize that before.
Does it help you to shift a shebang into your script?
Seems to be a SysV thing. I reproduced this behavior on HP-UX and AIX.
If there is no shebang, the execve() does not change the args, so they remain from the previous fork() i.e. are a copy from the invoking shell.
If i remove #! /bin/bash from the script other commands are getting affected. So, I need #! /bin/bash else my script fails for other reasons.
The reason i wrote the script is only to check if the script is running as a process or not.
What can be done ?
The shebang must be there!
$ cat myscript
#! /bin/bash
$ sleep 888
./myscript &
[1] 7164
$ pgrep -l myscript
7164 myscript