I haven't set up Spam Assassin, but I do run a lot of mailing lists. You
can eliminate 99% of all spam on mailing lists by doing two things:
1) Only members of the list can post to the list (I don't think you're
doing this, but not sure since I'm not on the other list being discussed)
2) You must answer an email to join the list (I think you already do this)
There are other things that can be done to get rid of the last 1%, but this
gets it down to very controlled levels for most mailing lists. If the list
admin wants to take this off line, I'd be happy to discuss it, and work
with him until the problem is resolved.
At 11:09 AM 1/30/2004, you wrote:
>Brion Vibber idly wondered,
> > > Can we install Spam Assassin software on the wikien-l mailing list?
> > > How about other mailing lists?
> > Do you know how to set it up properly as a machine-wide service?
>Are you kidding? All I know is VB and Java on Windows. The only service
>I know anything about takes place on Sundays at my church...
>But I know what e-mail filtering is: certain combinations of keywords
>tip off the software to flag a message as spam, and automatically moves
>it to another bin. You can either leave the suspected spam in the bin,
>or double-check it, or just have it automatically deleted.
>Wikitech-l mailing list
I'm getting really confused, here.
How can a squid wear boots, let alone get re-booted? And are we
preparing for a raid, or conducting one? What are we, jack-booted thugs
dealing out death in the night?
Now, a squid might wear a floppy if he was a SEAL...
But it sounds more like a bio-war facility with viruses and bugs, I just
hope you know what you're doing Jimbo!!
We're in the process of switching the mailing lists over to the new
servers. The offsite backup mail server was supposed to pick up the
slack while the DNS changes propagate, but apparently there've been
some troubles with it. Sorry for any inconvenience; hopefully
it'll all be working within a couple days at most for those having
-- brion vibber (brion @ pobox.com)
Just for information:
I got an access denied error when trying to reach the german lang. WP.
The requested URL could not be retrieved
While trying to retrieve the URL: http://de.wikipedia.org/
The following error was encountered:
Access control configuration prevents your request from being allowed at this
time. Please contact your service provider if you feel this is incorrect.
Your cache administrator is webmaster.
Generated Thu, 05 Feb 2004 10:49:41 GMT by localhost.localdomain
2 Minutes later it worked again.
Discussed on French Wikipedia:
Some wikipedians would like to have an automatic detection of "questionable
modifications". Questionable modifications, mean, modifications that have a
great chance to be a rookie's test or a vandalism. This automatic detection
may be done by check, for example:
* Article size reduction limit (or %): someone replace an article by "I rule
Wikipedia!" or cut half the article.
* Minimal size of new article: someone create an article with only "Hello
* Length of words or letters repetition in small articles:
There are many other criteria that can define "questionable modifications",
but this can handle a lot of case we run up against. Those questionable
modifications may be displayed in RC (for sysop or for all) with a different
mark (color, bold, italic or any visible marker). The goal is to give
information on witch articles have to be checked.
I'm setting up iptables filtering on the new machines to keep prying
eyes outside of local services. I've noticed a couple oddities:
Lots of mysterious UDP broadcasts from 184.108.40.206. This subnet is
listed on several spam blacklists; I've just dropped all connections
from that IP to keep it out of the logs.
Every 5 seconds there's a UDP broadcast to port 712 from 220.127.116.11
and 18.104.22.168. This port is apparently for "Topology Broadcast
Based on Reverse-Path Forwarding (TBRPF)"...? I've set it to drop and
I also saw an ICMP type 10 ("Router solicitation message"?) or two from
10.135.1.2, which is on the reserved 10/8 network. Either they're using
the reserved network for some sort of internal routing at this place,
or somebody's spoofing. I don't really like the idea of spoofing, since
some of our internal services are authenticated partially or wholly by
IP (such as memcached) and it'd be nice to be sure that something
claiming to be in our subnet really is.
-- brion vibber (brion @ pobox.com)
Just for fun, I tried updating the firmware so that our existing SCSI
card would support mirroring. That did not work, so hardware RAID
is off the table for now.
In this case, given the limitations of the card, hardware RAID would
have been less desirable anyway, perhaps, because the card only
supports 0 and 1, not 5. This would limit us to 36GB of disk space.
With software RAID, we can do RAID 5, so that our 3x36 can work as
So, I think that's what we should do.
All 9 machines ran the full battery of tests (not just the default
8 tests, but all eleven) for 24 hours without a single error on any
of the machines! ding!
Additionally, just for fun, I stuck in some RAM from my home machine
(G5), and it works just fine, too. This is a different brand.
Probably not all that important to know, except that with our
longstanding tradition of nightmarish RAM problems, it's nice to
already know of two brands that work perfectly fine.
There is one very important quirk that I need to report. The
dual opteron reports 4,096MB of ram, as expected.
However, the two Pentium IV servers with 4 physical gig of RAM in them
report only 3,584MB of RAM. I fiddled with the chips to confirm that
each of them really is 1 Gig, and that's not it. I also poked around
in the bios (in an uneducated manner), but saw nothing to fix.
Here's the setup... I haven't yet physically labelled the machines
with their names. This list starts with the Opteron and goes up in ip
numbers by one.
2U Opteron - 4 gig of ram
1 - 2
1 - 2
1 - 4
1 - 4
1 - 1
1 - 1
1 - 1
1 - 1
The idea is that for now, we'll have 2 squids, 2 webservers, and 1
database server. The other 4 machines with 1 gig of ram could be
pressed into service with or without ram upgrades, or can be used
as-is for auxiliary services. At least one is going to be the mail
server, and probably another will be the nameserver.
memtest86 will run overnight, and if (as I expect) all goes well, we
will be ready for production tomorrow. If the machines can only
recognize 3.5GB of RAM instead of 4, that isn't a joyful thing, but
it really doesn't matter a LOT, I don't think.