or possibly use a little script instead of tail.. the below seems to work, but I don't know how good an idea it would be to use it on -really- active logfiles. Consider it my quick and dirty friday night excercise ;-)
This assumes a file ./logfiles exists, in which the files to watch are listed (t1 t2 t3). Every 5 seconds it checks if the no. of lines has increased. If so, the new lines are logged. If a file suddenly has less lines, a log reset is assumed and the file is logged from line 1.
It's one job to launch (or kill), you can tweak the intervals (and pretty much anything, of course), and if a file gets reset with content ( echo "resetting log" > t1 ) it doesn't seem to miss the first line as tail tends to do.
Again, I have not truly tested this or anything, so please do and improve where needed, but who knows it could be useful.
Cheers
Wout
#!/bin/sh
FIRST=1
while true
do
if [ $FIRST -eq 1 ]
then
for i in $(
do
eval ${i}C=$(wc -l $i|awk '{print $1}')
done
FIRST=0
else
for i in $(
do
eval ${i}N=$(wc -l $i|awk '{print $1}')
eval NN=\$${i}N
eval CC=\$${i}C
if [ $NN -gt $CC ]
then
eval ${i}C=$((${i}C+1))
eval sed -n "\$${i}C,\${${i}N}p" $i | logger
eval ${i}C=\$${i}N
elif [ $NN -lt $CC ]
then
echo "$i decreased in size, assuming reset." | logger
eval sed -n "1,\${${i}N}p" $i | logger
eval ${i}C=\$${i}N
fi
done
fi
sleep 3
done
an engineer's aim in a discussion is not to persuade, but to clarify.
