log script

cmontr · 06-11-2009, 10:32 AM

Hello guys,

Does anyone have a sample script that if a log file gets greater then 10MB and deletes it?

Thanks

colucix · 06-11-2009, 10:37 AM

You have at least three possibilities:
1) use the find command with -size and -exec.
2) use the stat command with -c%s and if the size is greater than a certain value delete the file
3) add a rule to logrotate and specify the size option.

cmontr · 06-11-2009, 10:56 AM

This will be in Solaris box...by any chance do you have a sample one that I can modify?
Thanks

Quote:

Originally Posted by colucix

You have at least three possibilities:
1) use the find command with -size and -exec.
2) use the stat command with -c%s and if the size is greater than a certain value delete the file
3) add a rule to logrotate and specify the size option.

colucix · 06-11-2009, 11:09 AM

Code:

find . -name logfile -size +10000000c -exec echo rm {} \;

Note than on Solaris the -size option accepts either a number of 512-bytes blocks or the size in bytes (using the c suffix as in my example). The echo statement is just for testing purpose (remove it if you want to actually delete the file).

cmontr · 06-11-2009, 11:26 AM

This works, thank you very much.
find . -name test.log -size +10c -exec cat /dev/null >test.log {} \;

Is there a way to add a condition to it such as say:

if test.log > 10MB
then keep last 2MB logged
then delete the rest

Thanks much

Quote:

Originally Posted by colucix

Code:

find . -name logfile -size +10000000c -exec echo rm {} \;

Note than on Solaris the -size option accepts either a number of 512-bytes blocks or the size in bytes (using the c suffix as in my example). The echo statement is just for testing purpose (remove it if you want to actually delete the file).

colucix · 06-11-2009, 11:51 AM

Well.. in that case I'd use split and a simple script like this:

Code:

#!/bin/sh
size=`ls -l logfile | awk '{print $5}'`
if [ $size -gt 10000000 ]
then
  tail -r logfile > reversed_logfile
  split -b 2m reversed_logfile
  tail -r xaa > logfile
  rm reversed_logfile x??
fi

The tail -r command is to be sure you get the last 2 mb of the file, otherwise the last part of the original splitted file may contains only few kb.

cmontr · 06-11-2009, 01:02 PM

tried this script...it runs but something not working. Could you explain what xaa and x?? supposed to be? I modified few times but not getting results. Thanks again.

Quote:

Originally Posted by colucix

Well.. in that case I'd use split and a simple script like this:

Code:

#!/bin/sh
size=`ls -l logfile | awk '{print $5}'`
if [ $size -gt 10000000 ]
then
  tail -r logfile > reversed_logfile
  split -b 2m reversed_logfile
  tail -r xaa > logfile
  rm reversed_logfile x??
fi

The tail -r command is to be sure you get the last 2 mb of the file, otherwise the last part of the original splitted file may contains only few kb.

colucix · 06-12-2009, 02:19 AM

Well, xaa, xab, xac... are the default name of the pieces created by the split command. Go through the man page of these command is something is not clear or if some behaviour is different. I have tested it on a Solaris Sparc 5.8 and it works.

Anyway, please post the exact commands you tried and copy/paste the error messages. If you simply tell "something don't work".... you tell me nothing!