Nope, we’re not affected. KTHXBYE.
We’ll be shutting jabber.ccc.de down for at about 10:45 CET today since there are planned maintenance works at our colo’s power supply.
These might last a bit too long for our UPS to cope with the outage.
As mentioned in https://github.com/stpeter/manifesto, we changed the server’s config to require TLS for the server-to-server links.
This might (and possibly will) destroy communications with your friends at Google Talk, since up to now they (Google) do not support TLS on their s2s side of XMPP.
Discussion about this test day can also be found at the XMPP operators list: http://mail.jabber.org/mailman/listinfo/operators
This morning we temporarily disabled Roster Versioning (XEP-0237) since it wasn’t working and only threw client errors when requested.
Just in case you haven’t noticed: We’re experiencing quite a nasty bug (maybe, not proven yet) inside ejabberd which leads to a) ejabberd eating all of the available memory and b) afterwards having one thread run at 100% CPU. At this time, ejabberd stops forwarding packets and accepting new connections.
Just to keep you informed.
P.S.: We do have a Twitter account that might (!) have more up-to-date information than you can find in this blog: @jabberccc.de
Some hours ago we deployed new SSL certificates – those fingerprints can be found at the SSL page.
Luckily this fixed the S2S issues we had with some other servers, too.
And since some of use are on holiday and IPMI isn’t working this is about some other entities to help us more.
FYI: We might move to a different hardware this month, so the not-so-working remote console will be fixed.
Server is down, will need remote hands, this will take some time.
09:26 – we’re back.
DNS amplification attack, same issue as had happened yesterday.
Edit 2013-08-28, 15:45: At least the attack on jabber.ccc.de seems to have been ‘resolved’.
(FYI: This server was being attacked, too – don’t forget to have a look at @jabbercccde on Twitter for additional up-to-date information)