Full Recv-Queue
Summary
After a few hours, the bind will no longer respond to all LISTEN addresses. OS: netbsd-8, bind-9.10.7 and bind-9.10.8
Steps to reproduce
I can not generate the error.
What is the current bug behavior?
After some time, the bind will not respond to all listen addresses. netstat Shows:
udp 4899 0 139.18.25.33.53 . udp 0 0 139.18.25.34.53 .
Requests for IP IP 139.18.25.34 will be answered correctly. Requests for IP 139.18.25.33 will not be answered. netstat shows a full queue for incoming packets on this IP. It does not always affect this IP, it also happens that another of the listen addresses is affected. It only helps restart the named. After that, everything works correctly for a while.
What is the expected correct behavior?
expected is an empty Recv-Queue
Relevant configuration files
options { directory "/etc/namedb"; dnssec-enable yes; dnssec-validation auto; allow-recursion { any; };
rate-limit {
responses-per-second 25;
exempt-clients {127.0.0.0/8; 192.168.0.0/16; 172.16.0.0/12; 10.0.0.0/8; 139.18.0.0/16; 2001:638:902::0/48; };
};
listen-on { 139.18.25.33; 139.18.25.34; 192.88.99.1; };
listen-on-v6 { 2001:638:902:1::1; 2001:638:902:1::10; };};
Relevant logs and/or screenshots
Nothing found in the logfiles.