Please read this.
The server doesn't have enough RAM. Nothing is guaranteed to work properly. We're working on it.
On the one hand, I sympathize and wish only the best for blogs.perl.org. It's a huge improvement over the status quo, and a tough service to provide. On the other... you should be able to handle at least tens of requests per second on half a gig of RAM, and should simply slow down rather than throwing random errors if you can't. $^M might help here.
educated_foo wrote: you should be able to handle at least tens of requests per second on ...
This problem should be resolved in a week or two and in the meantime, it's very, very much "use at your own risk". It's been stated repeatedly that the site is not done and that people should expect problems. I don't know what more can be said aside from "there will be an announcement once this has been resolved".
Thanks for the response -- I'll be patient. Still, I'm surprised that MT is so resource-hungry and ungraceful when it fails. Clearly I don't do enough web development.
It's not MT per se - it's that the server is running close enough to the bone that operations like fork() fail sporadically, causing basic functions to error out. It seems to crop up most often in page rebuilding.
I blog about Perl.