Folks,
I have a large whitelist (380M) and this code is death when remove_entry
is called:
if ($addr =~ /^(.*)\|ip=cmd$/) {
# it doesn't have an IP attached.
# try to delete any per-IP entries for this addr as well.
# could be slow...
my $mailaddr = $1;
my @keys = grep { /^\Q${mailaddr}\E\|ip=(?:\d+\.\d+|none)$/ }
keys %{$self->{accum}};
foreach my $key (@keys) {
delete $self->{accum}->{$key};
delete $self->{accum}->{$key.'|totscore'};
}
It takes about 5 minutes to run and 900M of swap, which backs everything
else up behind it. Somebody tried to add about 50 blacklisted domains to
our company-wide SpamAssassin, and wiped us out for an hour.
I was thinking about taking a crack at this to rewrite the $addr store to
pack all of the IP lookups in a single key, something like this:
[EMAIL PROTECTED]|count cmd=0|10.1=9
[EMAIL PROTECTED]|totscore cmd=-5|10.1=215
etc. This requires the subclass to unpack the address and do some extra
packing/unpacking, which I didn't particularly like.
The other option was to maintain a seperate associated AWL which would
contain a list of IPs that were associated so one could easily delete the
keys without the grep having to run:
[EMAIL PROTECTED] 10.1,10.2,209.212,...
Any thoughts?
Cheers,
David.
-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills. Sign up for IBM's
Free Linux Tutorials. Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
