On 08/27/2013 04:05 PM, Mark Campbell wrote:
well, it's interesting, 'cause I didn't get any of the emails back until I noticed the sites working again... And I'd tried accessing from both work & home. perhaps a downed router between the two of us? Always glad to hear of high uptimes from linux devices ;)
It's possible the machine was just DoS-ed by a botnet trying to brute-force the the CMS password. It happens all the time, and since it's only a 1.2GHz ARM and WordPress is a bloated pile of excrutiatingly slow crap, it didn't take much to slow the machine down to the point where it was no longer up to the job of serving any valid requests. I have counter-measures in place for that sort of thing, but since the requests were coming from thousands of IPs, they were ineffective.
Since then I have added an iptables rule that simply tarpits all HTTP packets trying to access wp-login from everywhere except my own IP. Problem solved. :) And the long term load average on the machine dropped drastically as a consequence, too, according to my graphs.
As for the updates repo, I just assumed you'd do it in the same way that CentOS does with its repos. I suppose it works for them to keep the updates in a separate repo, maybe so that you can always fall back to a known-working release build. But ultimately, as long as I can get the updates as they are compiled, that's all I really care about.
I may change the way it hangs together later. When it happens I'll also setup up a proper mirrorlist and issue a new redsleeve-release package containing the updated .repo files.
Gordan _______________________________________________ users mailing list [email protected] http://lists.redsleeve.org/mailman/listinfo/users
