This reloader didn't close the openened file handle. Add a close. Can't
use `defer` because this is in a endless loop.
Signed-off-by: Miek Gieben <miek@miek.nl>
* plugin/file: respond correctly to IXFR message
Respond with a sing SOA record to an IXFR request if the SOA serials
match.
The added test fails on the current code with:
~~~
=== RUN TestIxfrResponse
--- FAIL: TestIxfrResponse (0.00s)
secondary_test.go:122: Expected answer section with single RR
FAIL
exit status 1
~~~
And obviously passes with the new code. This should cut down on the
weird number of zone transfers that I was seeing. At some point IXFR
support might be cool.
Fixes: #3176
Signed-off-by: Miek Gieben <miek@miek.nl>
* reuse code
Signed-off-by: Miek Gieben <miek@miek.nl>
* Sligtht tweaks
Signed-off-by: Miek Gieben <miek@miek.nl>
This can't be used in its current form; revert the entire PR.
Revert "fuzzing: allow setup function to be called (#3175)"
This reverts commit 62451fd3eb.
* fuzz: use gofuzz build tag instead of fuzz
Since go-fuzz does not support Go modules yet, vendor dependencies.
Signed-off-by: Muhammad Falak R Wani <falakreyaz@gmail.com>
* fuzz: avoid vendoring code for go-fuzz
Signed-off-by: Muhammad Falak R Wani <falakreyaz@gmail.com>
* fix external ns records
* use k8s service name for ns record
* update test, add func comment
* expand nsAddrs() test cases
* support local ipv6 ip
* use less confusing pod ip in test
* plugin/clouddns: tiny fixes for the README
Did a post-merge review. Fix a few typos.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Naming;
coredns -> CoreDNS
clouddns -> Cloud DNS
and italics then the plugin's name are mentioned.
Signed-off-by: Miek Gieben <miek@miek.nl>
This PR adds some content in clouddns plugin to mention about
the fact that privately hosted zone does not need to attach to
a VPC.
Also change PROJECT_NAME to PROJECTT_ID, and reformt the markdown
to replace tab with 4 whitespace.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
In the setup function use plugin.Error() to wrap the errors with the
plugin name. Because there isn't a separate setup() function this is
done for all returned errors.
Remove *upstream.Upstream from the New parameters as this is always set
and adjust the tests to account for this.
Signed-off-by: Miek Gieben <miek@miek.nl>
* fix switch order
* remove fallthough
* add test
* fix test
* distingush nxdomain/nodata for at least first subdomain of zone
* restore fallthough; reorder switch cases
document the environment option and some cleanups. Go over the code and
fix/tweak random bits here and there.
Condense a few lines here and there.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Add plugin for Azure DNS
Signed-off-by: darshanime <deathbullet@gmail.com>
* Rename AzureDNS plugin to Azure
Signed-off-by: darshanime <deathbullet@gmail.com>
* remove upstream from azure syntax
Signed-off-by: darshanime <deathbullet@gmail.com>
* Rename azure plugin block keynames
Signed-off-by: darshanime <deathbullet@gmail.com>
* Normalize zone name before lookup in zones
Signed-off-by: darshanime <deathbullet@gmail.com>
* Update import path for caddy
Signed-off-by: darshanime <deathbullet@gmail.com>
* normalize azure zone name only if required
Signed-off-by: darshanime <deathbullet@gmail.com>
* Add support for MX, SRV, TXT, records
Signed-off-by: darshanime <deathbullet@gmail.com>
* Add specs for new record types
Signed-off-by: darshanime <deathbullet@gmail.com>
* Use sequential updates for zones
Signed-off-by: darshanime <deathbullet@gmail.com>
* Add OWNERS file for azure plugin
Signed-off-by: darshanime <deathbullet@gmail.com>
* Rename imports for third party packages
Signed-off-by: darshanime <deathbullet@gmail.com>
* Capitalize values in README
Signed-off-by: darshanime <deathbullet@gmail.com>
* Shorten keys for azure plugin config
Signed-off-by: darshanime <deathbullet@gmail.com>
* Fixup readme for azure plugin
Signed-off-by: darshanime <deathbullet@gmail.com>
the current update frequency for the refresh loop in the route 53 plugin is hard-coded
to 1 minute. aws rate-limits the number of api requests so less frequent record refreshes
can help when reaching those limits depending upon your individual scenarios. this pull
adds a configuration option to the route53 plugin to adjust the refresh frequency.
thanks for getting my last pull released so quickly. this is the last local change that
i have been running and would love to get it contributed back to the project.
Signed-off-by: Matt Kulka <mkulka@parchment.com>
Various fixes to make things less flaky:
* kubernetes: put klog.SetOutput in the setup function, not in the init
function to see if that helps
* file: make z.Expired a boolean instead of a pointer to a boolean
* test: fix TestSecondaryZoneTransfer test, which wasn't actually
testing in the right way. It's more right now, but may still be racy
(race introduced because a file's lazy loading of zones)
Signed-off-by: Miek Gieben <miek@miek.nl>
* plugin/hosts: create inline map in setup
The inline map wasn't create in the setup.go fuction leading to a crash,
which is masked by a recover (but leads to a SERVFAIL, and not logging
the request).
Various other simplifications.
host plugin could use some test that actually cover these edgecases.
Signed-off-by: Miek Gieben <miek@miek.nl>
* PR review changes
Signed-off-by: Miek Gieben <miek@miek.nl>
without the paging parameter set, it will default to 100 records per request. with large enough zones and
potentially multiple coredns daemons configured to pull from route 53, this can quickly add up and reach
aws global api rate limits (5/sec per root account). increasing paging to max can help reduce the number
of requests needed to pull records for a zone without no down side that i am aware of.
this helps issue #2353, but probably is not a complete fix.
Signed-off-by: Matt Kulka <mkulka@parchment.com>
Move github.com/openzipkin/zipkin-go-opentracing => github.com/openzipkin-contrib/zipkin-go-opentracing
and run `go mod tidy`
Fixes 3069
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
I don't believe this is actually needed (anymore). The:
os.Stderr = os.Stdout
is a crazy hack that def. needs to go.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Fix corefile usage
* plugin/file: allow README.md testing
Allow readme testing for the file plugin and fix bugs that where found:
* the reader wasn't reset when re-reading the same io.reader for a
different origin.
Signed-off-by: Miek Gieben <miek@miek.nl>
* Update test/example_test.go
Co-Authored-By: Michael Grosser <development@stp-ip.net>
