6066570359 | ||
---|---|---|
README.MD | ||
cache.go | ||
godns.conf | ||
godns_test.go | ||
handler.go | ||
hosts.go | ||
main.go | ||
resolver.go | ||
server.go | ||
settings.go |
README.MD
GODNS
A simple and fast dns cache server written by go.
Similar as dnsmasq ,but support some difference features:
-
Keep hosts records in redis instead of the local file /etc/hosts
-
Atuo-Reload when hosts configuration changed. (Yes,dnsmasq need restart)
-
Cache records save in memory or redis configurable
Install & Running
-
Install
$ go get github.com/kenshinx/godns
-
Build
$ cd $GOPATH/src/github.com/kenshinx/godns $ go build -o godns *.go
-
Running
$ sudo ./godns -c godns.conf
-
Use
$ sudo vi /etc/resolv.conf nameserver 127.0.0.1
Configuration
All the configuration on godns.conf
a TOML formating config file.
More about Toml :https://github.com/mojombo/toml
resolv.conf
Upstream server can be configuration by change file from somewhere other that "/etc/resolv.conf"
[resolv]
resolv-file = "/etc/resolv.conf"
If multi namerserver
set at resolv.conf, the upsteam server will try in order of up to botton
cache
Only the local memory storage backend implemented now. The redis backend is in todo list
[cache]
backend = "memory"
expire = 600 # default expire time 10 minutes
maxcount = 100000
hosts
Force resolv domain to assigned ip, support two types hosts configuration:
- locale hosts file
- remote redis hosts
hosts file
can be assigned at godns.conf,default : /etc/hosts
[hosts]
host-file = "/etc/hosts"
redis hosts
This is a espeical requirment in our system. Must maintain a gloab hosts configuration, and support update the hosts record from other remote server. so "redis-hosts" is be supported, and will query the reids when each dns request reached.
The hosts record is organized with redis hash map. and the key the map is configired.
[hosts]
redis-key = "godns:hosts"
Benchmak
Debug close
$ go test -bench=.
testing: warning: no tests to run
PASS
BenchmarkDig-4 10000 202732 ns/op
ok _/Users/kenshin/workspace/gogo/godns 2.489s
The result : 4032 queries/per second
The enviroment of test:
MacBook Air
-
CPU:
Inter Core i5 1.7G
Double cores -
MEM:
8G
TODO
- The redis cache backend
- Update ttl