kubeconfig arg. is not flowed through to other command invocations

[sync-by-unito[bot]]

Use of the the --kubeconfig arg. does not correctly flow through to e.g. kubectl invocations. To see how this can cause an error do the following:

1. delete your own dev. cluster (or otherwise setup your normal /home/user/.kube/config file to be stale
2. create a cluster with a custom kube config file somewhere specific (non-default location)
3. try px deploy --kubeconfig /path/to/your/kube/config

However, {{KUBECONFIG=/path/to/your/kube/config px deploy }} appears to work.

┆Issue is synchronized with this Jira Bug by Unito

[victor-timofei]

Is anyone working on this? If not, I could pick it up.

[htroisi]

@victor-timofei - We would be happy to accept another contribution from you! 😀

@zasgar is in the process of completely rewriting the CLI so he may already have this bug fixed. However, would you be interested in contributing a new feature to Pixie? I just filed a feature request to add Node labels to Pixie's metadata (https://github.com/pixie-io/pixie/issues/580). Another contributor recently added Pod labels to Pixie's metadata, so after checking out that PR (linked in the above issue), it should be fairly straightforward to add node labels.

[victor-timofei]

@htroisi yeah sure, I'll start working on the #580. 😄

[htroisi]

@victor-timofei - Awesome! You can find the directions for setting up the developer environment here. There's also some recent discussion about the dev environment in the #pixienaut channel in our community slack.

[victor-timofei]

I did already set up a dev environment, if I encounter any issues with the environment(particularly with running the tests) I'll try searching the #pixienaut history or ask a question.

Thanks 😀