* Fully deprecate NO_RELOAD
Signed-off-by: Xiao An <hac@zju.edu.cn>
* Fully deprecate TIMEOUT
Signed-off-by: Xiao An <hac@zju.edu.cn>
* Update CI tests to adapt to the deprecation of TIMEOUT
Signed-off-by: Xiao An <hac@zju.edu.cn>
* Add documents for directive transfer in plugin auto
Signed-off-by: Xiao An <hac@zju.edu.cn>
* plugin/chaos: add default list of authors
Add a owners_generate.go that generates a Owners variables for use in
the chaos plugin.
Add a default list of authors in the authors.bind CH zone. When doing a
query this now returns:
~~~ sh
% dig authors.bind TXT CH
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 5456
;; flags: qr rd; QUERY: 1, ANSWER: 22, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;authors.bind. CH TXT
;; ANSWER SECTION:
authors.bind. 0 CH TXT "bradbeam"
authors.bind. 0 CH TXT "chrisohaver"
authors.bind. 0 CH TXT "dilyevsky"
authors.bind. 0 CH TXT "ekleiner"
authors.bind. 0 CH TXT "fastest963"
authors.bind. 0 CH TXT "fturib"
authors.bind. 0 CH TXT "greenpau"
authors.bind. 0 CH TXT "grobie"
authors.bind. 0 CH TXT "inigohu"
authors.bind. 0 CH TXT "isolus"
authors.bind. 0 CH TXT "johnbelamaric"
authors.bind. 0 CH TXT "miekg"
authors.bind. 0 CH TXT "nchrisdk"
authors.bind. 0 CH TXT "nitisht"
authors.bind. 0 CH TXT "pmoroney"
authors.bind. 0 CH TXT "rajansandeep"
authors.bind. 0 CH TXT "rdrozhdzh"
authors.bind. 0 CH TXT "rtreffer"
authors.bind. 0 CH TXT "stp-ip"
authors.bind. 0 CH TXT "superq"
authors.bind. 0 CH TXT "varyoo"
authors.bind. 0 CH TXT "yongtang"
~~~
This was hard to do previously as we didn't hardcode this in the source,
but now with OWNERS files we can just generate this list.
Privacy wise this isn't worse than being listed in OWNERS file in the
first place. And it's a nice hat tip to the people making CoreDNS
better.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Sticklet bot comments
Signed-off-by: Miek Gieben <miek@miek.nl>
* [plugin/route53]: Do not return NXDOMAIN where it should be NODATA.
Signed-off-by: Dmitry Ilyevskiy <dmitry.ilyevskiy@getcruise.com>
* Fix bad merge.
Signed-off-by: Dmitry Ilyevskiy <dmitry.ilyevskiy@getcruise.com>
* plugin/cancel: add context cancelation plugin
Per review comments on #2704, move this into a plugin that gets called.
Add the most minimal plugin, tests and documenation.
Signed-off-by: Miek Gieben <miek@miek.nl>
* plugin/cache: add timeout option
review feedback: add option to set custom timeout.
Signed-off-by: Miek Gieben <miek@miek.nl>
* spelling
Signed-off-by: Miek Gieben <miek@miek.nl>
* Remove context.Context from request.Request
This removes the context from request.Request and makes all the changes
in the code to make it compile again. It's all mechanical. It did
unearth some weirdness in that the context was kept in handler structs
which may cause havoc with concurrently handling of requests.
Fixes#2721
Signed-off-by: Miek Gieben <miek@miek.nl>
* Make test compile
Signed-off-by: Miek Gieben <miek@miek.nl>
Add another test case for impossible DNS messages which should not be
cached. This adds a check for a message that denies its own existence.
Fixes#2724.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Don't double report metrics on error
When there is an error use a different function to report the metrics,
in case the plugin chain handled the request the metrics are already
reported.
Fixes: #2717
Signed-off-by: Miek Gieben <miek@miek.nl>
* Compile again
Signed-off-by: Miek Gieben <miek@miek.nl>
* more
Signed-off-by: Miek Gieben <miek@miek.nl>
* Remove server addr from the context
This was added twice, just leave the server which also holds the
address.
Conflicts with #2719 but should be easy to fix.
Signed-off-by: Miek Gieben <miek@miek.nl>
* doesn't need server context
Signed-off-by: Miek Gieben <miek@miek.nl>
The server handles this case no need to also do it in the log plugin.
Means DefaultErrorFunc can be private to the dnsserver and is now
renamed to just errorFunc
Fixes: #2715
Signed-off-by: Miek Gieben <miek@miek.nl>
* Add a GaugeVec for enabled plugins monitoring.
Signed-off-by: Jiacheng Xu <xjcmaxwellcjx@gmail.com>
* Add server label and zone label for enable_plugin matric.
* Add a test for PluginEnabled metric
* Add description for enabledPlugin metric.
* Change the description for the enabledPlugin metric.
* Reset the enabledPlugin metric when restart the server.
* Add the bug session for enabledPlugin metric.
* Remove the resolveTCPAddr
* plugin/kubernetes: add ready function
Add ready function as the health function is now gone.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Fix readme
Signed-off-by: Miek Gieben <miek@miek.nl>
* plugin/grpc: New gRPC plugin
* some changes after the first review:
- remove healthcheck. gRPC already has this implicitly implemented
- some naming and stetic changes
- fix some comments
- other minor fixes
* plugin/grpc: New gRPC plugin
* some changes after the first review:
- remove healthcheck. gRPC already has this implicitly implemented
- some naming and stetic changes
- fix some comments
- other minor fixes
* add OWNERS file and change plugin order
* remove Rcode checker
* Remove etcd unit testing from travis
Remove these from the travis unit testing. (to speed up travis)
See #2671
Signed-off-by: Miek Gieben <miek@miek.nl>
* Add build tag
Signed-off-by: Miek Gieben <miek@miek.nl>
* Take 3: to get coredns/proxy pulled in
This is need temporary because it will point to the fake 'deprecated'
plugin anyway.
Signed-off-by: Miek Gieben <miek@miek.nl>
* update proxy again
Signed-off-by: Miek Gieben <miek@miek.nl>
* plugin/health: remove ability to poll other plugins
This mechanism defeats the purpose any plugin (mostly) caching can still
be alive, we can probably forward queries still. Don't poll plugins,
just tell the world we're up and running.
It was only actually used in kubernetes; and there specifically would
mean any network hiccup would NACK the entire server health.
Fixes: #2534
Signed-off-by: Miek Gieben <miek@miek.nl>
* update docs based on feedback
Signed-off-by: Miek Gieben <miek@miek.nl>
Add a ready plugin that allows plugin to signal when they are ready.
Once a plugin is ready it is not queried again.
This uses same mechanism as the health plugin: each plugin needs to
implement an interface.
Implement readines for the *erratic* plugin to aid in testing.
Add README.md and tests moduled after the health plugin; which will be
relegated to just providing process health. In similar vein to health
this is a process wide setting.
With this Corefile:
~~~
. {
erratic
whoami
ready
}
bla {
erratic
whoami
}
~~~
ready will lead to:
~~~ sh
% curl localhost:8181/ready
% dig @localhost -p 1053 mx example.org
% curl localhost:8181/ready
OK%
~~~
Meanwhile CoreDNS logs:
~~~
.:1053
bla.:1053
2019-02-26T20:59:07.137Z [INFO] CoreDNS-1.3.1
2019-02-26T20:59:07.137Z [INFO] linux/amd64, go1.11.4,
CoreDNS-1.3.1
linux/amd64, go1.11.4,
2019-02-26T20:59:11.415Z [INFO] plugin/ready: Still waiting on: "erratic"
2019-02-26T20:59:13.510Z [INFO] plugin/ready: Still waiting on: "erratic"
~~~
*ready* can be used in multiple server blocks and will do the right
thing; query all those plugins from all server blocks for readiness.
This does a similar thing to the prometheus plugin.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Remove internal loop detection
I can't actually think of a situation where we can create an internal
loop. Sure externally triggered cycles can happen, but this is where the
*loop* plugin comes in that detects those.
Fixes#2602
Signed-off-by: Miek Gieben <miek@miek.nl>
* Remove test
Signed-off-by: Miek Gieben <miek@miek.nl>
* Move *proxy* to external
move the proxy plugin into coredns/proxy and remove it as a default
plugin. Link the proxy to deprecated in plugin.cfg
coredns/proxy doesn't compile because of the vendoring :(
Signed-off-by: Miek Gieben <miek@miek.nl>
* Add github.com/coredns/proxy
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
