4

This works but the way I did this is kind of silly. Is there a better way to do this? 

for e in $(ipcs | awk '{print $2}'); do
    [[ "$e" = "Semaphore" ]] && break
    echo $e
    ipcrm shm $e
done

    echo $e outside for loop
for e in $(ipcs | awk '{print $2}'); do
    [[ "$e" = "Message" ]] && break
    echo $e
    ipcrm -s $e
done

echo
exit 0

This is what ipcs looks like for me when I run it. 

$ ipcs

------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status      
0x00000000 262145     bob        600        393216     2          dest         
0x00000000 2523138    bob        600        393216     2          dest         
0x00000000 2555907    bob        600        393216     2          dest         
0x00000000 3375108    bob        600        998400     2          dest         
0x00000000 3440645    bob        666        40         1                       

------ Semaphore Arrays --------
key        semid      owner      perms      nsems     
0x000005b1 262146     bob        600        6         

------ Message Queues --------
key        msqid      owner      perms      used-bytes   messages

I need to run these two commands on a regular basis. 

ipcrm -s $(ipcs | grep bob | awk '{printf "%s ",$2}')
ipcrm  shm $(ipcs | grep bob | awk '{printf "%s ",$2}')

So I was thinking I could do something like this. 

if [ `$(ipcs | grep Shared | awk '{print $2}')` == "Shared"]
ipcrm  shm $(ipcs | grep bob | awk '{printf "%s ",$2}')

I want to do that first behavior until $2 equals Semaphore. 

if [ `$(ipcs | grep Semaphore | awk '{print $2}')` == "Semaphore"]
ipcrm -s $(ipcs | grep bob | awk '{printf "%s ",$2}'

So to summarize I want the first if block to run after I see "Shared". Then I want the second if block after I see "Semaphore".

Improve this question
1
  • You can use a case statement instead of two if statements on separate loops. Anyway I don't see what you are trying to achieve clearly enough. Commented Oct 10, 2014 at 8:24

3 Answers 3

Reset to default
6

If I understand corretly, you want to run the command ipcrm shm <ids> for all Shared Memory Segments of the user bob. Then the command ipcrm -s <ids> for all Semaphore Arrays of user bob.

To accomplish that use the following commands (you don't have to loop in a script):

For the Shared Memory Segments part:

ipcrm shm $(ipcs -m | awk '$3=="bob" {printf "%s ",$2}')

For the Semaphore Array part:

ipcrm -s $(ipcs -s | awk '$3=="bob" {printf "%s ",$2}')

Explanation:

From the ipcs manpage:

   -m     shared memory segments
   -s     semaphore arrays

The awk part prints the ids only when the 3rd field is bob.

Improve this answer
2
  • how does it tell the difference between the Semaphore Arrays and Shared Memory Segments block? You haven't done anything to account for that. As I said above I need to use both those commands on a regular basis so I would like a script. I don't wanna have to dig through my bash history every time or type it out every time. Commented Oct 10, 2014 at 9:26
  • 1
    @cokedude the -s and -m flag. It tells ipcs to only show Semaphore Arrays (-s) or Shared Memory Segments (-m). And of course you can put those two commands in a script (would be a small script). Commented Oct 10, 2014 at 9:33
2

Maybe...

ipcs | sed -n '
    s/[^ ]*  *//
    /^Messages/q
    /^Semaphores/cshift
    /  *bob .*/!d;s///
    / /!s/./ipcrm $1 &/p
'| sh -s -- shm \-s

That deletes lines that do not contain the string bob as the third space delimited field or that do not have in their second field Messages/Semaphores.

It inserts the string ipcrm $1 <field 2> for those lines that remain. It quits input when matching Messages and it replaces a Semaphores match w/ shift.

sed's output is interpreted by a shell process with two positionals shm/-s. So when sed says shift that shell stops running the command ipcrm shm <field2> and starts running -s in shms place.

I guess if you wanted a pure shell solution this comes close:

set -f; IFS='
'; for l in $(ipcs); 
do IFS=\ ;set -- $l
case "$1:$2:${3#bob}" in
(-*:Sh*) a=shm;; 
(-*:Se*) a=-s;; 
(-*:Me*) break 2;;
(*:*:) ipcrm "$a" "$2";;
esac; done
Improve this answer
-2

As slm suggested on your other question, I suggest doing this with a while loop rather than a for loop: 

looking_at=
ipcs | while read key id owner other
do
      # Keep track of which section we’re in.
      if [ "$id" = "Shared" ]               # $id is the second field; i.e., $2.
      then
            looking_at=shm
      elif [ "$id" = "Semaphore" ]
      then
            looking_at=sem
      elif [ "$id" = "Message" ]
      then
            break
      fi
      if [ "$owner" = "bob" ]
      then
            if [ "$looking_at" = "shm" ]
            then
                  ipcrm shm "$id"
            elif [ "$looking_at" = "sem" ]
            then
                  ipcrm -s "$id"
            fi
      fi
done

This reads the output of the ipcs command one line at a time, breaking out the first three fields as key, id, and owner.  As the comment says, we use the looking_at variable to keep track of which section we’re in.  Then, on every line that contains bob in the third field, we use looking_at to determine which ipcrm option to use.

Improve this answer
0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.