jamesoff.net




26
Sep
2006
3

I’ve had enough

I’ve decided I’ve had enough of blbl.org. Last night my server that runs it (along with all of my other stuff) had some kind of hardware fault, and I realised that I didn’t really want to pay for this machine just for blbl.org to rag it until the disks fall out and the assorted important bits melt. My father always used to say there’s nothing worse than having your important bits melt.

Actually he didn’t, but if I was to ask him about this theoretical event his response would definitely be on the negative side.

What it comes down to is this: blbl.org sucks quite a bit of power out of my server in both memory use and CPU time. (The bandwidth is probably negligible, but the worry of some blog spammer getting his knickers in a twist and performing some kind of DDoS is always at the back of my mind.) It hammers the disk (the main database tables contain 18 million and 6 million rows respectively, and that’s not taking into account the tables used to accumulate the data people submit) and generally slows everything down. I’m getting fed up of having to field questions like “why is your gallery down?” because of it.

On top of all of that, I seem to end up spending more time than I’d like looking after the whole (admitedly shaky) set up. If blbl.org was represented on paper, it would be a Heath Robinson machine. Frankly SpamKarma2 does a pretty damn good job of trapping spam by itself, WordPress has anti-spam stuff in it and there are plenty of other plugins which do a decent job too.

Looking at the stats from rbldnsd(8), I was fielding about 400 queries every five minutes (i.e. just over one per second) for the IP blacklist, and I was getting about a 10% hit rate. I guess that’s not so bad, but I’ve no idea how much of the comments posted on all blogs ever (or all blogs that might use blbl.org) is spam. If it’s 10%, then I guess I was doing ok, but I suspect that it’s probably more than that. Based on the flood of submissions (which often ran to 200/sec), I’d say it’s not that effective anyway.

The stats for the URL blacklist are worse – roughly 100 queries every five minutes and less than four hits in the same period.

So, I’ve shut it off. blbl.org is now firewalled off to save my webserver from even having to attempt to serve requests. In about a week, the scripts that generate the blacklist data for rbldnsd will decide nothing should be listed any more. I’ll leave that running for now because otherwise it will have an impact on people’s blogs. 200+ submissions a second tells me I could potentially screw up a lot of blogs :) I certainly won’t be doing anything like returning a positive hit for all queries and saying “well you should have noticed that blbl.org was shutting down!”

In due time, I’ll shut down rbldnsd too, but that’ll be a while yet. If you’re using the SpamKarma2 plugin for WordPress, or any other RBL lookup plugin on any blog software at all, please remove blbl.org from the lookup list (bl.blbl.org and uri-bl.blbl.org).

I would say it’s been fun, but I’m not sure that applies :) It’s certainly been interesting from many perspectives. I’ve had to figure out how to run an RBL, write scripts (and rewrite) them to handle large volumes of submissions, process large quantities of data, and so on. I never even got round to writing a proper website for it (not that you can look now, since I’ve blocked it). I hope that in the time it’s been running I’ve helped stop at least a little bit of spam, and stuck two fingers up at the people who go around filling people’s blogs with unnecessary “comments” about poker, transexuals, lesbians and bestiality. There are plenty of people who write their own blogs about that. Actually, I think that’s called myspace or something :fry:

In the time it’s been running, especially since I posted last about the detrimental effect blbl.org was having on my server (and thus all my other sites), I’ve had a couple of offers from people for more hosting. I’ve decided not to take them up in the end (although I am very grateful) because leaving this project is as much about my personal time as anything else. When I had more free time (like when I was a student) I happily created many complex projects for myself, but nowadays I want to keep things simpler because I’m so busy.

If anyone wants to take over blbl.org (in its entirely – management and hosting) then I’d be more than happy to hand the reigns over to a suitable person (or people).

I think this is officially my longest blog post ever, so I’d better stop here before I ruin my average.


Permalink | Posted in RBL Spam internets 



4
May
2006
1

Uh oh

I can’t help but worry about blbl.org after reading stories like this:

http://b.oooom.net/261


Permalink | Posted in RBL Spam internets 



28
Mar
2006
5

blbl.org kills my stuff

Unfortunately, blbl.org (my BLog Blocking List) is generating a lot of load on my poor colo’d machine, and is causing general slowness and at times outright failures of my other websites.

Also unfortunately for blbl.org, I like my other websites being up more than I like blbl.org being up, so this means one of several things:

  • I shut it down
  • I move it to another machine that can handle it
  • I change it so it generates less load

(more…)





26
Jun
2005
0

SK2 RBL Plugin Update

See http://www.grooblehonk.co.uk/sk2/rbl-plugin/ for details.

  • Don’t look up same host more than once per post
  • Fixed URIs not getting looked up properly

The downloadable file has been updated.


Permalink | Posted in Coding RBL 



10
Jun
2005
1

RBL Plugin available for SK2

If you want to try out the new improved RBL plugin for Spam Karma 2, have a look here. It works fine for me so I’d like some feed back from other people now :)


Permalink | Posted in Coding RBL Spam 



26
May
2005
0

SK2 RBL Plugin Update

After a good evening’s coding last night I’ve got a working version of the Spam Karma 2 RBL plugin which looks IPs and URIs up in different RBLs, and can also submit back to blbl.org when it finds a spammy comment.

Still got some testing to do and also tidying up the code, adding options etc to do. Hopefully I’ll put a version for testing up shortly.

Related to the above, I’ve changed blbl.org to list IPs under bl.blbl.org and URIs under uri-bl.blbl.org. I shall be giving the blbl.org site a bit of a facelift soon.


Permalink | Posted in Coding RBL Spam 



24
May
2005
0

Spam Karma 2

Wow, SK2 rocks.

Dr Dave has invited me to write the auto RBL-submission code for it which I shall hopefully be doing this week sometime. I just looked at the todo list for 2.0 and spotted that auto-submission is one of the few outstanding items left, so I need to get my skates on :)

It also seems MediaWikis (like bmotion.net’s) are targets for spammers too. A quick bit of investigating this morning shows that MediaWiki actually sets a rel on external links to stop search engines following them off the wiki so spamming is useless in that sense. It just clogs up the blog and can generate clickthoughs by users. There is some anti-spam stuff available and I shall look at adding RBL support to it too.


Permalink | Posted in Coding RBL Spam 



18
Mar
2005
1

Free comment spamtrap day!

For a limited period only all comments posted on here will be tagged spam and added to the RBL.


Permalink | Posted in RBL Spamtrap 



21
Feb
2005
0

blbl.org spamtrap

trap.blbl.org is a wordpress spamtrap for blbl.org. All comments posted there will be ignored but the poster’s IP will be added to the RBL automatically. Let’s hope some spammers pick it up :)


Permalink | Posted in RBL Spam 



6
Jan
2005
0

Auto-RBL submission for SK

Having played with Spam Karma a bit, I’ve now added (experimental) support for auto-submitting IPs and URLs to the RBL. They’re auto-submitted if a spam scores the maximum amount and is auto-deleted.

I’ve also changed the regular expression used for pulling out URLs from the comment so that it strips trailing slashes off the URL, although it still needs some work as it will sometimes pull bits of HTML out of the comment too.

If you want to try it (which would be much appreciated both for testing my code and for testing the RBL submission and updating) please hit the RBL homepage for instructions.


Permalink | Posted in Coding RBL Spam