Set klog's logtostderr flag

[gjtempleton]

1. Why is this pull request needed and what does it do?

The update of client-go to v10.0.0 has moved K8s logging from glog to klog.

This change removes the previous non-obvious flag setting behaviour used for glog and then ensures that klog, the replacement for glog has the relevant flag set. See klog code for the reason why this has to happen: https://github.com/coredns/coredns/blob/92836cc6f9aaf47a349993271ddae273d8e7d25d/vendor/k8s.io/klog/klog.go

Note that simply updating to the version of klog updating this flag to default to true rather than false was tested and still caused crashes.

Tested locally:

kubectl logs -f -n kube-system core-dns-8ff86fc8c-s2hng
.:53
2019-02-05T20:52:29.173Z [INFO] CoreDNS-1.3.1
2019-02-05T20:52:29.173Z [INFO] linux/amd64, go1.11.4, 66623192-dirty
CoreDNS-1.3.1
linux/amd64, go1.11.4, 66623192-dirty
2019-02-05T20:52:29.173Z [INFO] plugin/reload: Running configuration MD5 = a03e157fe14ec2d191bddec59dc496fc
E0205 20:53:15.981439       1 streamwatcher.go:109] Unable to decode an event from the watch stream: http2: server sent GOAWAY and closed the connection; LastStreamID=15, ErrCode=NO_ERROR, debug=""
E0205 20:53:15.981458       1 streamwatcher.go:109] Unable to decode an event from the watch stream: http2: server sent GOAWAY and closed the connection; LastStreamID=15, ErrCode=NO_ERROR, debug=""
E0205 20:53:15.981458       1 streamwatcher.go:109] Unable to decode an event from the watch stream: http2: server sent GOAWAY and closed the connection; LastStreamID=15, ErrCode=NO_ERROR, debug=""
E0205 20:53:15.981470       1 streamwatcher.go:109] Unable to decode an event from the watch stream: http2: server sent GOAWAY and closed the connection; LastStreamID=15, ErrCode=NO_ERROR, debug=""
E0205 20:53:15.981849       1 reflector.go:251] github.com/coredns/coredns/plugin/kubernetes/controller.go:322: Failed to watch *v1.Namespace: Get https://10.254.0.1:443/api/v1/namespaces?resourceVersion=343852&timeout=5m46s&timeoutSeconds=346&watch=true: dial tcp 10.254.0.1:443: connect: connection refused
E0205 20:53:15.981851       1 reflector.go:251] github.com/coredns/coredns/plugin/kubernetes/controller.go:320: Failed to watch *v1.Pod: Get https://10.254.0.1:443/api/v1/pods?resourceVersion=347745&timeout=5m19s&timeoutSeconds=319&watch=true: dial tcp 10.254.0.1:443: connect: connection refused
E0205 20:53:15.981851       1 reflector.go:251] github.com/coredns/coredns/plugin/kubernetes/controller.go:317: Failed to watch *v1.Endpoints: Get https://10.254.0.1:443/api/v1/endpoints?resourceVersion=347877&timeout=5m29s&timeoutSeconds=329&watch=true: dial tcp 10.254.0.1:443: connect: connection refused
E0205 20:53:15.981876       1 reflector.go:251] github.com/coredns/coredns/plugin/kubernetes/controller.go:315: Failed to watch *v1.Service: Get https://10.254.0.1:443/api/v1/services?resourceVersion=343852&timeout=8m26s&timeoutSeconds=506&watch=true: dial tcp 10.254.0.1:443: connect: connection refused

No restarts of pods observed.

2. Which issues (if any) are related?

#2464 , supersedes the sticking plaster attempted by #2525

3. Which documentation changes (if any) need to be made?

Noted as a bug fix for a regression in v1.3.1 only.

[corbot]

Thank you for your contribution. I've just checked the OWNERS files to find a suitable reviewer. This search was successful and I've asked fturib (via plugin/kubernetes/OWNERS) for a review.

If you have questions or suggestions for this bot, please file an issue against the miekg/dreck repository.

The bot understands the commands that are listed here.

[codecov-io]

Codecov Report

Merging #2529 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #2529      +/-   ##
==========================================
+ Coverage   55.37%   55.38%   +0.01%     
==========================================
  Files         204      204              
  Lines       10346    10349       +3     
==========================================
+ Hits         5729     5732       +3     
- Misses       4196     4197       +1     
+ Partials      421      420       -1

Impacted Files	Coverage Δ
plugin/kubernetes/setup.go	70.46% <100%> (+0.46%)	⬆️
plugin/forward/connect.go	80.59% <0%> (-4.48%)	⬇️
plugin/route53/route53.go	85.18% <0%> (+2.77%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 92836cc...5dd3935. Read the comment docs.

[miekg]

I though klog was better than glog? Remove the 'ugh' at least (glog is almost evil on how it interacts with its environment)

[gjtempleton]

Yeah, it is a significant step forward from glog and looks to be making progress on continuing to clean up a lot of the minor pain points, have cleared up the comment.

Think it's worth still keeping there given the current behaviour, at least as used by the k8s libraries, is still to log to disk.