I'm trying to extract stuff from the STEEM database using HTTP RPC API calls on steemd. It works nicely most of the time, but sometimes I receive garbled responses, and I've already crashed my steemd a couple of times.
"Garbled" means I've seen a "200 Assumed OK" status line in the response, while the body contained several leading 0-bytes, another "200 OK" line, a content-length header, some JSON, and what looks like the remains of responses from previous requests. I've also seen responses with an obviously incomplete JSON body.
I've been able to reproduce the problem by hammering the node with several thousand requests using a simple perl script.
Ideas anyone?