[foaf-dev] FOAF site - unscheduled downtime

[Dan Brickley]

On Fri, Sep 17, 2010 at 12:20 AM, Dan Brickley <danbri at danbri.org> wrote:
> On Thu, Sep 16, 2010 at 11:16 PM, Damian Steer <pldms at mac.com> wrote:
>> This has happened to me, too. Kingsley's right, try a reboot. It's not the same as stopping the instance, the local storage remains,
>
> Thanks Kingsley, Damian, everyone! Yes, an AWS console reboot brought
> it back from the very edge of oblivion.
>
> Something still isn't quite right - the load jumps straight to "
> 22:10:38 up 2 min,  1 user,  load average: 27.31, 9.31, 3.33". Also we
> realised that the anti-spam tools on the wiki are inadequate and it
> has been filled with rubbish again; not clear yet whether the issues
> are related.

I managed to shut down the Web server (apache2), and took a look at
the bigger log files,

find . -name \*.log -exec ls -lh {} \;  | grep G

Sure enough, it was the mediawiki installation, which again is under
sustained automatic spam attack. If they'd shown a bit more restraint
it would have gone on longer without my noticing.

I've disabled wiki.foaf-project.org for now. When it comes back up,
I'll make it 'read only' except for those of us in FOAF club. Quite
how we define that I'm open to suggestions. But first rule of FOAF
club - don't be a spammer! Everything else is a nice extra. I think
it's time to eat our own dogfood here, so I'll be revisiting
http://danbri.org/words/2009/10/25/504 and looking around for other
sources of trusted openids and webids too.

I think things are stable for now. Sorry for the wiki downtime,

cheers,

Dan