Script help required - calling the guru's

[Simon R Wootton]

Hopefully nothing too complex.

I have a directory containing many thousands of small files. Each day, about 20,000 files are added to this directory. Each night, a cron task runs to delete the files in this directory that are over 7 days old.

Instead of deleting the files, I need to archive them, but I'm concious of inode limits within filesystems.

Therefore, I'm looking for a script that will take the 20,000 or so files from one day, compress them into 1 file (using gzip?) and name that file as the date being compressed (eg. 140704.Z)

Anyone fancy a crack ??

Help will be rewarded generously!

Regards
Simon

[Geoff Wild · Accepted solution]

How about something like:

find /20000directory -mtime +7 -print >/tmp/filelist

TARFILE=`/bin/date +%y%m%d%H%M%S`.tar
FILES=`cat /tmp/filelist`

tar cvf /archdir/$TARFILE $FILES

gzip /20000directory/$TARFILE



Rgds...Geoff

[harry d brown jr]

Modify your script to use GNU's tar (http://hpux.cs.utah.edu/hppd/hpux/Gnu/tar-1.14/).

When using GNU's tar, use the --gzip. Once you have archived the file, delete it.

live free or die
harry

[Steve Steel]

Hi

find dir -mtime +7|xargs shar|gzip > $(/bin/date +%y%m%d%H%M%S).gz

Puts them in a compressed shared archive

Then remove them

Steve steel

[Kenneth Platz]

One-liner:

find /mydir -mtime +7 | tar cvf - | gzip > /mydir/$(date +%d%m%y).tar.gz

Enjoy :-)

[Kenneth Platz]

Correction. That should be:

tar cvf - $( find /mydir -mtime +7 ) | gzip > /mydir/$( date +%d%m%y ).tar.gz