I recently wrote a log zipping script specific to the environment I currently work in.  We don't remove logs from the servers (yet – that will change, but has not so far).  The partitions on which the logs live are quite large.  The servers' lifespan is short enough before rebuild or replace to make leaving logs in place possible. In many cases no compressing of logs is needed.  On some servers the traffic and the application do create files that must be compressed to keep free space on the partition acceptable.  

We had an ad hoc script to do this in detail.  The complexity in it was in trying to work around not touching live logs that a process was writing to, tweaking and tuning the regular expressions to miss those active files  and still catch the ones wanted.   Lots of hand work. 

#! /bin/bash

###
### version: 20100414.01
### author:  dsm
###
### Note - "rm -f $i >> $LOG 2>&1" for deleting testing
###        log files is commented out in 2 places
###        to prevent mistake - uncomment if you really
###        want to delete files.  Until
###        uncommented, files that would have been deleted
###        will by logged but not actually
###        deleted from the filesystem...
###

### VARIABLES

# file to alter retention for individual servers
# create the file via something like
#   export MYHOST=`hostname -s`
#   echo "3" > /etc/$MYHOST.AGE.CTCT_gzip_and_cleanup_logs
# This would change the AGE variable to "3"
# this resolves having different files to
# propagate for individual servers or losing the settings
# when a new version is copied out...

MYSELF=`hostname -s`
MYSELF_AGE_FILE="/etc/gzip/$MYSELF.AGE.gzip_and_cleanup_logs"

DATE=`date +%y%m%d%H%M%S`
LOG="/var/logs/gzip_and_cleanup.${DATE}"
date > $LOG # initialize log
if [[ -f $MYSELF_AGE_FILE ]]; then
    AGE=`cat $MYSELF_AGE_FILE`
else
    AGE=7 # files must be older than $AGE days to be gzipped
fi
# debug
#echo "age set to: $AGE"

LSOF="/usr/sbin/lsof"  #location of lsof binary...
if [[ -f $LSOF ]]; then
    echo "found $LSOF, continuing..." >> $LOG
else
    echo "FAILED TO FIND LSOF AT:  $LSOF - EXITING" >> $LOG
    echo "this will have to be corrected before this script will run" >> $LOG
    exit
fi

GZIP="/usr/bin/gzip"
if [[ -f $GZIP ]]; then
    echo "found $GZIP, continuing..." >> $LOG
else
    echo "FAILED TO FIND gzip AT:  $GZIP - EXITING" >> $LOG
    echo "this will have to be corrected before this script will run" >> $LOG
    exit
fi

#!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
TESTING="false" # this enables the delele function
              # NOT TO BE ENABLED ON PRODUCTION MACHINES!
KEEP=30 # days to keep, older will be deleted if TESTING set to "true"
#!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

### MAIN

#
# If TESTING:  Delete logfiles older than $KEEP
#
if [[ $TESTING == "true" ]]; then
    export LOGDIR=/var/logs
    cd ${LOGDIR}
    date >> $LOG
    echo "" >> $LOG
    echo "-----------------------------------">> $LOG
    echo "start $LOGDIR DELETE:  TESTING is $TESTING" >> $LOG
    echo "-----------------------------------" >> $LOG
    for i in `ls --color=no | grep -i log | egrep -v .gz$`
    do
        if [[ `find . -name ${i} -mtime +${KEEP}` ]]; then
    	    if [[ `$LSOF $i` ]]; then
	    	echo "found $i older than $KEEP - but found to be opened by a process - NOT deleting" >> $LOG
	    else
                echo "found $i older than $KEEP:  deleting" >> $LOG
                # rm -f $i >> $LOG 2>&1
	    fi
        else
            echo "$i found to be newer than $KEEP:  keeping this file" >> $LOG
        fi
    done
    export LOGDIR=/var/logs/http
    cd ${LOGDIR}
    echo "" >> $LOG
    echo "-----------------------------------">> $LOG
    echo "start $LOGDIR DELETE:  TESTING is $TESTING" >> $LOG
    echo "-----------------------------------" >> $LOG
    for i in `ls --color=no | grep -i log | egrep -v .gz$`
    do
        if [[ `find . -name ${i} -mtime +${KEEP}` ]]; then
 	    if [[ `$LSOF $i` ]]; then
                echo "found $i older than $KEEP - but found to be opened by a process - NOT deleting" >> $LOG
            else
                echo "found $i older than $KEEP:  deleting" >> $LOG
                # rm -f $i >> $LOG 2>&1
	    fi
        else
            echo "$i found to be newer than $KEEP:  keeping this file" >> $LOG
        fi
    done
else
    echo "TESTING set to false:  $TESTING:  NOT deleting any log files..." >> $LOG
fi

#
# Begin cleanup with application server logs
#

export LOGDIR=/var/logs
cd ${LOGDIR} >> $LOG 2>&1
date >> $LOG
echo "" >> $LOG
echo "-----------------------------------">> $LOG
echo "start $LOGDIR cleanup" >> $LOG
echo "-----------------------------------" >> $LOG
echo "" >> $LOG
for i in `ls --color=no | grep -i log | egrep -v .gz$`
do
    if [[ `$LSOF ${i}` ]]; then
        echo "" >> $LOG
	echo "found $i OPEN - not handling" >> $LOG
	echo "" >> $LOG
    else
        if [[ `find . -name ${i} -mtime +${AGE}` ]]; then
	    if [[ -f ${i}.gz ]]; then
	        echo "found gzipped file of the same name:  ${i}.gz - not gzipping ${i}..." >> $LOG
	    else
	        echo "gzipping $i as it is older than $AGE..." >> $LOG
	        $GZIP $i >> $LOG 2>&1
	    fi
	else
	    echo "$i is not open but is newer than $AGE, so NOT being gzipped" >> $LOG
        fi
    fi
done
echo "completed cleanup of application logs..." >> $LOG

#
# Continue cleanup with web server logs
#

export LOGDIR=/var/logs/http
if [[ -d $LOGDIR ]]; then # these may not exist on some apps...
    cd ${LOGDIR} >> $LOG 2>&1
    echo "" >> $LOG
    echo "-----------------------------------">> $LOG
    echo "start $LOGDIR cleanup" >> $LOG
    echo "-----------------------------------" >> $LOG
    echo "" >> $LOG
    for i in `ls --color=no | grep -i log | egrep -v .gz$ | egrep -v error_log$ | egrep -v access_log$`
    do
        if [[ `$LSOF ${i}` ]]; then
	    echo "" >> $LOG
            echo "found $i OPEN - not handling" >> $LOG
	    echo "" >> $LOG
        else
	    if [[ `find . -name ${i} -mtime +${AGE}` ]]; then
		if [[ -f ${i}.gz ]]; then
	            echo "found gzipped file of the same name:  ${i}.gz - not gzipping ${i}..." >> $LOG
	        else
                    echo "gzipping $i as it is older than $AGE..." >> $LOG
                    $GZIP $i >> $LOG 2>&1
		fi
	    else
                echo "$i is not open but is newer than $AGE, so NOT being gzipped" >> $LOG
            fi
        fi
    done
    echo "completed cleanup of apache logs..." >> $LOG
else
    echo "-----------------------------------">> $LOG
    echo "start $LOGDIR cleanup" >> $LOG
    echo "-----------------------------------" >> $LOG
    echo "" >> $LOG
    echo "??? failed to find $LOGDIR - PROBLEM???" >> $LOG
    echo "" >> $LOG
fi

#
# Cleanup logs once processed...
#

# these logs are post-processing - no need to check for lsof and no need to retain unzipped...

export LOGDIR=/var/logs/app
if [[ -d $LOGDIR ]]; then
    cd ${LOGDIR} >> $LOG 2>&1
    echo "" >> $LOG
    echo "-----------------------------------">> $LOG
    echo "start app $LOGDIR cleanup" >> $LOG
    echo "-----------------------------------" >> $LOG
    echo "" >> $LOG
    for i in `ls --color=no | grep open0`
    do
	echo "looking through $LOGDIR/${i}/processed..." >> $LOG
        echo "" >> $LOG
        cd $LOGDIR/${i}/processed
        for j in `ls --color=no | egrep -v .gz$`
	do
	    if [[ -f ${j}.gz ]]; then
                echo "found gzipped file of the same name:  ${j}.gz - not gzipping ${j}..." >> $LOG
            else
                echo "gzipping $j app log..." >> $LOG
                    $GZIP $j >> $LOG 2>&1
            fi
	done
        echo "COMPLETED $LOGDIR/${i}/processed..." >> $LOG
	echo "" >> $LOG
    done
    echo "" >> $LOG
    echo "Processing ${LOGDIR}/processed app logs..." >> $LOG
    echo "" >> $LOG
    cd ${LOGDIR}/processed
    for k in `ls --color=no | egrep -v .gz$`
    do
        if [[ -f ${k}.gz ]]; then
            echo "found gzipped file of the same name:  ${k}.gz - not gzipping ${k}..." >> $LOG
        else
            echo "gzipping $k app log..." >> $LOG
            $GZIP $k >> $LOG 2>&1
        fi
    done
    echo "completed app logs cleanup..." >> $LOG
else
    echo "" >> $LOG
    echo "failed to find $LOGDIR - skipping cleanup of application logs..." >> $LOG
    echo "" >> $LOG
fi 

#### Changelog ####
dsm - edited for broad issue

 

He had a huge loving heart, and leaves an overwhelming hole behind him.  He was there for me without any question or reservation and I hope he begins to know how much he is loved and how much he is missed.  How proud I am of him.  He was a very good man.

I'm leaving tomorrow after six weeks out with her.  That's good – I miss being home, I miss my family in MA.  I am so glad my daughter is better enough that I can see leaving.

But leaving isn't easy, either. 

–doug

My daughter has visually more healthy bone marrow than she had 36 days ago when she came into the hospital in crisis.   She has at least managed to knock the cancer back and rebuild a bone marrow that is not diseased.  More detailed tests will look for any trace at all of cancerous marrow.  I'm hoping there will be none to find at all.  That would be remission, the first step toward a life outside of a hospital room.

–doug