Fabian Groffen
Fabian Groffen
Do you have a test-case for this? Port and protocol should be checked (based on saddr), so I don't expect this to be necessary at all.
I think this is because we don't use blocks or frames, which is probably a shared problem for lzo, lz4 and snappy
Ok, can you describe the traffic? do you have a lot of connections that stay alive, or are they all connect+disconnects? I'd really like a stacktrace, but I guess I'll...
Ok, some questions: 1. did you compile this carbon-c-relay yourself or did it come from some distro repo? 2. if you compiled it yourself, would you be willing to re-compile...
Hey sorry for not responding, your backtrace is interesting, and I meant to look at this, but life is a bit hectic around here at the moment. Glad to hear...
this issue might actually be related to concurrency
You can't at the moment. I wonder why you feel this is necessary though?
hmmm, and how would you like to control the amount of connections to use per destination?
Do you have stats about queue sizes on the other targets?
Can you zoom in, e.g. from 13:20 to 13:40 to show the queue sizes? Not that this is in any way more of a fact, but what you're looking at...