UNIX for Dummies Questions & Answers

Hi All,
I am new to unix programming. I am trying for a requirement and the requirement goes like this.....

I have a test folder. Which tracks log files. After certain time, the log file is getting overwritten by another file (randomly as the time interval is not periodic). I need to preserve all the log files in a different folder with the timestamp before it gets overwritten.

Ex:
The log file contents and size varies with time
test/test1.log - 07/22/2011 13:01:11 ; 2 MB
test/test1.log - 07/22/2011 13:27:12 ; 2 MB
test/test1.log - 07/22/2011 14:27:09 ; 1.5 MB

I need to create a archive folder and preserve all these 3 files with their time stamp.

archive/test1.log_20110722130111
archive/test1.log_20110722132712
archive/test1.log_20110722142709


Any help on this is highly appreciated..

-Saravana

Assuming that the flavour of UNIX you are using has a stat command, then something like this could be used to create your filename:

Code:

log=/var/log/foo.log
dt=$(stat -c %Y $log |awk '{print  strftime( "%Y%m%d%H%M%S", $1 )}')  #make the datestamp portion in YYYYMMDDHHMMSS format
archive=/archive/log/foo_$dt.log

The stat command generates the epoch timestamp for the named file and the awk script is used to format it into something useful. The %y format supported by stat isn't very filename friendly.

If you are using FreeBSD, then the stat command format is a wee bit different. On the 6.2 BSD system I have it's: stat -f %a file-name

There is a caution in the man page for stat on Linux that states that a shell may implement an internal stat command that will override the binary in /bin

Thanks Agama...This helped a lot...
thank you so much....