I have a program that reads through a log file (from a firewall
actually) and generates a few statistics from it. Problem is that is uses a
ton of memory. I had to add a lot (500MB) of swap to a Linux box with 128
MB to begin with. BTW - Nothing else is running.
I commented out everything and added lines back one at a time until I got to
where things went bad. It was a command that does a split. Here's the
uncommented code that causes me to run out of memory - nothing fancy, just a
loop that reads and does a split on the input line:
#!/usr/bin/perl
#
# Program to count which firewall rules were applied how many times
#
#
open ( FWLOG, "./logfile.010626");
while ( $LINE = <FWLOG> ) {
chomp ($LINE);
@fw_log = split (/;/, $LINE);
}
The file, logfile.010626 is big, about 2.3 million records, but the records
are not that long. Is new memory allocated for each instance of $LINE or
for @fw_log? If so, is there a way I can make it reuse the same memory?
Or, is this just the way Perl does IO?
Thanks in advance!
Rob Blader
Naval Surface Warfare Center
(540)653-7270
[EMAIL PROTECTED]
