July 2002 - WLUG - wlug.mailman3.com

Re: [Wlug] Antispam software (was spam with header)

by Andy Stewart

On Tuesday 02 July 2002 3:18 pm, Karl Hiramoto wrote: > This would be an excellent meeting topic! I agree. If someone is willing to present this topic at a future WLUG meeting, let me know! Later, Andy -- Andy Stewart, Founder Worcester Linux Users' Group Worcester, MA USA http://www.wlug.org

July 3, 2002

Re: [Wlug] Antispam software (was spam with header)

by Karl Hiramoto

This would be an excellent meeting topic! On Tue, 2 Jul 2002, Theo Van Dinter wrote: > On Sun, Jun 30, 2002 at 10:38:24PM -0400, Karl Hiramoto wrote: > > Do you have a project created for this software? Distribute it? I > > have a spam problem @ work. Looking into solutions. I looked at maps > > rbl http://mail-abuse.org/rbl/ not sure on using it in a production > > environment though (where the most spam goes to) > > No, since it's mainly SpamAssassin with a little bit of procmail > frosting and a small perl script on top. I was planning to just make it > all available with the perpetually unfinished article I'm working on. > (It doesn't help that there were 2.5 new versions of SA released since > I started writing it ... <grrrr> ;) ) > > > My main worry is someone sending e-mail to sales(a)company.com and it > > gettting rejected, and comany.com loses a sale. Is this an issue? > > Well, that's the difference between a filter like SpamAssassin and > a blacklist like the RBL. Personally, I use several blacklists for > open-relays at the SMTP level, since it's unusual to have a false-positive > from an open-relay test (depending on the test of course). > > For a company, I would use a filter (like SA) and just let it mark up > likely spam. You can then leave it up to the users to decide what they > want to do with it. (SA adds headers indicating how likely a message > is to be spam (the score/number of hits), and what tests matched a > given message.) > > Anything else that makes it through to delivery goes through SpamAssassin. > Spams gets sorted into a seperate mail folder, and I then go through it > and report the spammers, etc. There's some information about what I do > for general filtering at: > > http://www.kluge.net/mailfiltering/ > > -- ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø Karl Hiramoto <karl(a)hiramoto.org> Work: 978-425-2090 ext 25 Cell: 508-517-4819 Personal web page: http://karl.hiramoto.org/ Zoop Productions: http://www.zoop.org/ KTEQ Rapid City: http://www.kteq.org/ AOL IM ID = KarlH420 Yahoo_IM = karl_hiramoto ¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø It's not so hard to lift yourself by your bootstraps once you're off the ground. -- Daniel B. Luten

July 2, 2002

Re: [Wlug] tmpfs

by Brian J. Conway

> --bind has made it into the mount man page, but the tmpfs is not > documented in the man pages anywhere I can think to look. Does anyone > know of a reason that it should not be used? I'm thinking of using it > for some locking files and pid stamps -- all stuff that should not > persist across a reboot. Tmpfs has some decent information in the kernel documentation. A lot of newer distros now mount /dev/shm as tmpfs, and it's basically a resizable ramdisk that will grow, shrink, or even swap out to disk as needed. I doubt that locking files are going to take up much space that you'd need back for memory, but on the other hand you probably don't need extra fast access to them either. ;-) Brian J. Conway bconway(a)wpi.edu "Any man who can drive safely while kissing a pretty girl is simply not giving the kiss the attention it deserves." - Albert Einstein

July 2, 2002

tmpfs

by Bill Mills-Curran

When researching the Reiserfs a while back, I discovered the tmpfs file system and the --bind option to mount. Both of these are nice tools to have around. --bind has made it into the mount man page, but the tmpfs is not documented in the man pages anywhere I can think to look. Does anyone know of a reason that it should not be used? I'm thinking of using it for some locking files and pid stamps -- all stuff that should not persist across a reboot. TIA, Bill

July 2, 2002

Re: [Wlug] Antispam software (was spam with header)

by Theo Van Dinter

On Sun, Jun 30, 2002 at 10:38:24PM -0400, Karl Hiramoto wrote: > Do you have a project created for this software? Distribute it? I > have a spam problem @ work. Looking into solutions. I looked at maps > rbl http://mail-abuse.org/rbl/ not sure on using it in a production > environment though (where the most spam goes to) No, since it's mainly SpamAssassin with a little bit of procmail frosting and a small perl script on top. I was planning to just make it all available with the perpetually unfinished article I'm working on. (It doesn't help that there were 2.5 new versions of SA released since I started writing it ... <grrrr> ;) ) > My main worry is someone sending e-mail to sales(a)company.com and it > gettting rejected, and comany.com loses a sale. Is this an issue? Well, that's the difference between a filter like SpamAssassin and a blacklist like the RBL. Personally, I use several blacklists for open-relays at the SMTP level, since it's unusual to have a false-positive from an open-relay test (depending on the test of course). For a company, I would use a filter (like SA) and just let it mark up likely spam. You can then leave it up to the users to decide what they want to do with it. (SA adds headers indicating how likely a message is to be spam (the score/number of hits), and what tests matched a given message.) Anything else that makes it through to delivery goes through SpamAssassin. Spams gets sorted into a seperate mail folder, and I then go through it and report the spammers, etc. There's some information about what I do for general filtering at: http://www.kluge.net/mailfiltering/ -- Randomly Generated Tagline: "And that's the success of Windows-it's mediocre, but it's easy." - Linus Torvalds

July 2, 2002