Jeff Lill

I just saw this crash again. This time I didn't press any special keys like HOME,... ``` root@neon-node-agent-c2dr5:/# cd /test root@neon-node-agent-c2dr5:/test# chmod 770 test.sh root@neon-node-agent-c2dr5:/test# cd .. root@neon-node-agent-c2dr5:/# echo /test/test.sh...

I can repo this fairly consistently by: 1. Exec'ing into a container to bash 2 Pressing the **UP ARROW** key a few times Doing the same in **kubectl** directly works...

I think we'll need to connect to the pod directly via a `WebSocket` and have **WinTTY** communicate directly with that. I believe we may also have a similar problem with...

- [ ] `ClusterProxy:` - [x] @jefflill: Reset namespaces - [ ] @jefflill: Reset CRI-O on nodes - [x] reset container registries - [ ] remove non built-in container images...

Delete namespace isn't working how I expected: * **default** namespace cannot be deleted, so we'll need to explicitly remove its contents * **deleting everything** in a namespace **doesn't always work**....

I've implemented the critical features and have added some todos to the code for the remaining ones and I'm going to move this issue back to the **Release Backlog** pipeline....

Moving this to the **product backlog**. This implementation is good enough for now.

@marcusbooyah: I'm not sure that making the domain a list really makes sense. I think that one domain should be enough for users as far as our UX goes. Having...

I've reported this to Azure: https://github.com/Azure/azure-sdk-for-net/issues/28914

The cluster is reporting being **partially configured** I investigated and `clusterLogin.SetupDetails.SetupPending` is still set to `true` for this cluster. I'll need to look into this further,