Hi Rasmus On Tue, Dec 15, 2015 at 03:20:05PM +0100, Rasmus Edgar wrote: > We started noticing 1s+ latency problems on clients resolving using the > vmware guest at a load around 6000 qps. > > Test setup: > > 1 x x86_64 vmware guest on Esx 5.5 > 8xVCPU > 8G RAM > vmxnet3 10Gb virtual interface > CentOS 7.1 > Bind 9.10.3 resolver > > 1xIBM x86_64 physical machine > 24xCPU cores > 16G ram > 1Gb interface > CentOS 7.1 > Bind 9.10.3 resolver
Generally VMs are not a proper place to test query performance, but anyway let's see. When named starts, it logs messages like this: 15-Dec-2015 20:08:00.245 found 8 CPUs, using 8 worker threads 15-Dec-2015 20:08:00.245 using 7 UDP listeners per interface 15-Dec-2015 20:08:00.245 using up to 4096 sockets Please will you paste what you get in your configuration? > ./dnsperf -f inet -s <redacted ip> -d queryfile-example-10million-201202 -l > 30 -q 15000 Why are you passing the -q argument to dnsperf? A high receive queue means your named process is not able to read off the socket fast enough. Try to find out where it's busy, using a profiler such as perf. Is it query logging or doing some other highly verbose logging that keeps it blocked in disk IO? Mukund
signature.asc
Description: PGP signature
_______________________________________________ Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users