Running 9.6.1-P1 on CentOS 5 with 8 CPU cores. We're anycasting our DNS
service
address. We observe the UDP Recv-Q get as high as 109540 and 2-4 second
response times. The simplest fix is to simply restart the bind process. When
the response time is slow on the DNS service address, I can do a query against
another address on the server, and the response time is normal, so it's
definitely correlated to DNS service address. The number of queries during the
restart didn't change, so obviously the way the bind process handles it is
different. Is there any way to further debug this behavior? Is this a bug, and
has it been fixed in newer versions? Thanks for any tips and insights.
_______________________________________________
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
