Re:Getting started with Tmux
by hardwaresofton on Sunday September 16, 2018 @08:15PM
(#18002133)
Thanks for the tip! Usually since I've just run a `kubectl get pods` or whatever I use the keyboard navigation to go up a few lines, ctrl+a and use the emacs bindings to yank...
I just downloaded and installed tmux-fingers with tpm, but it's doing a pretty terrible job of picking up on the right things... I'm assuming i need to put in the predefined regexes that would match pod names/ lines of the get output?
Re:Getting started with Tmux
by __blockcipher__ on Friday September 21, 2018 @09:30PM
(#18043959)
It works for the output of kubectl get all, not kubectl get po unfortunately. [This is because without the resource name as a prefix, there’s no general regexes that will work for everyone’s situation] As a hack I sometimes end up doing kubectl get all | grep ‘pod/‘, which is a little inelegant. You can add your own regexes (see the readme of tmux-fingers, you do @fingers-pattern-0 your_regex) and then reload the config
Realistically I suppose it’s only marginally faster than kubectl get po -o name and then manually copying.
For any project I work on frequently I set up aliases that wrap it anyway. As you already know kubectl doesn’t allow regex matching so I just do something like: kubectl delete pod “$(kubectl get po -l app=MY_APP_SELECTOR)”, etc. naturally your deploy needs granular selectors otherwise you need to further pipe to grep or awk
Kinda hard to explain, but it’ll make sense if you look at the gh repo.
[It’s always driven my insane how kubernetes names its pods. They still haven’t implemented regex matching - there’s an issue on github related to that that went stale. For now I use a combjnation of selectors and old school unix regexes (awk, grep, etc) when I’m passing around kubectl commands that I need to work rgardless of whether the container has restarted or not]