I'm wondering if increasing the backlog is not going to make the problem worse for you. You're machines can't cope already, and with that setting the load is going to be increased even more.

Currently what doesn't fit in backlog doesn't get processed.

On Thu, 25 Jan 2018 at 20:54 Christian Krbusek <christian@ph3x.at> wrote:

On 25-01-2018 20:33, Conrad Rockenhaus wrote:
> Have you checked your kernel socket backlog limit? (net.core.somaxconn)

Thanks for the hint. This is set to 128, a quick google revealed
settings
from torservers.net where it´s set to 20480 for a fast connection. I´ll
give
that a try and see how it works out.

Cheers,
Christian
_______________________________________________
tor-relays mailing list
tor-relays@lists.torproject.org
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays