Tero Marttila
Tero Marttila
The `api.endpoint` should only be used by the `pharos-cluster` tool itself, and not for the worker host kubelet -> master host kube-apiserver connections, so the scenario you describe should work,...
I'm guessing that this should work with `docker-compose run -T` to disable the tty allocation, for but whatever reason that also causes `pharos-cluster kubeconfig` to just hang without any output......
These will cause issues if trying to use `pharos-cluster reset` to switch `network.provider` without rebooting in between... for the calico -> weave case, the weave pods will be left in...
Same thing with weave... and switching from `weave` -> `calico` without rebooting leaves you with an unholy combination of having both weave and calico interfaces/routes/rules simultaneously, because calico doesn't validate...
Also just to confirm that it's the same jetty `Response` being called from the two threads: the `_fields: instance of org.eclipse.jetty.http.HttpFields(id=13856)` matches across both threads ``` Breakpoint hit: "thread=task-24", org.eclipse.jetty.server.Response.addHeader(),...
> Looking at the code, the [`ResponseBodyEmitter`](https://github.com/spring-projects/spring-framework/blob/v5.2.5.RELEASE/spring-webmvc/src/main/java/org/springframework/web/servlet/mvc/method/annotation/ResponseBodyEmitter.java) created by the [`ResponseBodyEmitterReturnValueHandler`](https://github.com/spring-projects/spring-framework/blob/v5.2.5.RELEASE/spring-webmvc/src/main/java/org/springframework/web/servlet/mvc/method/annotation/ResponseBodyEmitterReturnValueHandler.java)/[`ReactiveTypeHandler`](https://github.com/spring-projects/spring-framework/blob/v5.2.5.RELEASE/spring-webmvc/src/main/java/org/springframework/web/servlet/mvc/method/annotation/ReactiveTypeHandler.java) should not be writing any HTTP response headers outside of the [`extendResponse`](https://github.com/spring-projects/spring-framework/blob/v5.2.5.RELEASE/spring-webmvc/src/main/java/org/springframework/web/servlet/mvc/method/annotation/ResponseBodyEmitterReturnValueHandler.java#L170) called in the servlet thread, but the...
> Never mind, I see now. Yes, the [`ReactiveTypeHandler`](https://github.com/spring-projects/spring-framework/blob/v5.2.5.RELEASE/spring-webmvc/src/main/java/org/springframework/web/servlet/mvc/method/annotation/ReactiveTypeHandler.java#L137) adapts the `Flux
The problematic spring backend in our case did not need to override any Cache-Control headers, so we were able to deploy the `headerWriterFilter.setShouldWriteHeadersEagerly(true)` workaround. We can confirm that we are...
This appears to still be broken on Ubuntu xenial / ganeti 2.15.2-3, I'm unable to recreate the secondary LVM/DRBD devices on a node that failed, was reinstalled and readded to...
Full backtrace for the `resize: true` bug when the terminal width is too small for a column: ``` NoMethodError: undefined method `rindex' for nil:NilClass /home/kontena/kontena/kontena/cli/vendor/bundle/ruby/2.3.0/gems/verse-0.5.0/lib/verse/truncation.rb:71:in `truncate' /home/kontena/kontena/kontena/cli/vendor/bundle/ruby/2.3.0/gems/verse-0.5.0/lib/verse/truncation.rb:36:in `truncate' /home/kontena/kontena/kontena/cli/vendor/bundle/ruby/2.3.0/gems/verse-0.5.0/lib/verse.rb:47:in `truncate'...