|
|
# ? Jul 13, 2021 18:53 |
|
|
# ? May 3, 2024 05:45 |
|
Ansible Adams posted:yaml is fine it's really not as bad as old people make it out to be
|
# ? Jul 13, 2021 19:05 |
|
carry on then posted:it's really not as bad as old people make it out to be nothing is as bad as i make it out to be, but it's still bad
|
# ? Jul 13, 2021 19:14 |
|
Ansible Adams posted:yaml is fine
|
# ? Jul 13, 2021 19:18 |
|
shaggar and rotor were right
|
# ? Jul 13, 2021 19:37 |
|
yaml is like someone saw json and thought "this rules too much, let me make a shittier version"
|
# ? Jul 13, 2021 20:42 |
|
akadajet posted:yaml is like someone saw json and thought "this rules too much, let me make a shittier version" "You know what json needs? Comments. Also as long as i'm doing new stuff, here's a bunch of terrible choices."
|
# ? Jul 13, 2021 20:44 |
|
you, a moron: `kubectl exec -it my-lovely-pod-0 -- /bin/bash` me, a genius: `oc rsh my-cool-pod-0`
|
# ? Jul 13, 2021 22:04 |
|
you sayin a nete made these kubers??
|
# ? Jul 13, 2021 22:09 |
|
Lol at writing configuration files in XML
|
# ? Jul 13, 2021 23:04 |
|
abigserve posted:Lol at writing configuration files in XML it supports comments and it's not yaml
|
# ? Jul 13, 2021 23:16 |
|
kuberneteeznutz
|
# ? Jul 14, 2021 00:20 |
|
carry on then posted:you, a moron: `kubectl exec -it my-lovely-pod-0 -- /bin/bash` my beef with openshift is that they seem to feel the need to NIH like loving everything and entirely control the product. i know dealing with k8s oss governance can kinda suck but goddamn you don't need to replace the ENTIRE user-facing part of the ecosystem (unless you intentionally want to do this to get lockin and consulting $$$ where you're the one billing all the consultants, which given IBM is... yeah). this poo poo probably could have been a kubectl plugin dealing with the idiosyncrasies of EKS versus GCP versus roll-your-own is enough of its own workload but at least the UX part is the same, and i don't need to know about whatever OpenShift's replacement tooling or CRs are, esp since the OpenShift users seem to be less familiar with what they're doing and can't translate their questions to vanilla k8s terms at all. kubectl and a lot of the prominent stock tooling aren't so bad that they desperately need a replacement; they usually have enough people working with them that the UX pain gets hammered out over time
|
# ? Jul 14, 2021 01:54 |
|
dont know what anything in this thread is about
|
# ? Jul 14, 2021 01:56 |
|
hifi posted:dont know what anything in this thread is about janitoring in the cloud
|
# ? Jul 14, 2021 02:14 |
|
what happens if you just -i or -t but not both
|
# ? Jul 14, 2021 02:16 |
|
CMYK BLYAT! posted:my beef with openshift is that they seem to feel the need to NIH like loving everything and entirely control the product. i know dealing with k8s oss governance can kinda suck but goddamn you don't need to replace the ENTIRE user-facing part of the ecosystem (unless you intentionally want to do this to get lockin and consulting $$$ where you're the one billing all the consultants, which given IBM is... yeah). this poo poo probably could have been a kubectl plugin i'm sorry that someone is daring to improve on the kube user experience without first becoming a google employee i guess...?
|
# ? Jul 14, 2021 03:59 |
|
cool av posted:what happens if you just -i or -t but not both sorry but you have to dot the i's AND cross the t's
|
# ? Jul 14, 2021 04:09 |
|
carry on then posted:you, a moron: `kubectl exec -it my-lovely-pod-0 -- /bin/bash` aliasing kubectl to k is canon. also its pronounced cube-cuddle
|
# ? Jul 14, 2021 04:41 |
|
does anyone use nomad. I think they invented their own yaml-like
|
# ? Jul 14, 2021 05:00 |
|
Hed posted:they invented their own yaml-like so just a general comment here, if you're at work and someone says they're writing their own configuration file format you should respond by setting up a 1-on-1 meeting with them and breaking all their fingers.
|
# ? Jul 14, 2021 05:04 |
|
Hed posted:does anyone use nomad. I think they invented their own yaml-like nomad job files and terraform use the same config lang. its… okay
|
# ? Jul 14, 2021 06:10 |
|
you know what is worse than yaml? helm charts. here, write some loving yaml but also with go templating. loving hell
|
# ? Jul 14, 2021 06:12 |
|
ngl, what I'm starting to like even more than YAML is HCL2
|
# ? Jul 14, 2021 17:47 |
|
Bored Online posted:you know what is worse than yaml? helm charts. here, write some loving yaml but also with go templating. loving hell yah as someone who maintains a chart i can affirm that Helm is in many ways quite poo poo. really wish they had a proper kustomize integration so all the many many feature requests i get for "add this new deployment feature into the deployment" could be handled with "this doesn't need a loop or advanced template functionality, it is just stuffing a YAML blob into a defined place in another YAML blob, which is exactly what kustomize does, use that instead" but their solution for that is "uh you can run a bash script that invokes kustomize" which is poo poo. it's useful for de-duplicating poo poo for our 8 mostly-identical Services but other than that it's a hot mess. also someone submitted a PR the other day that'd let you put a template inside values.yaml and render it inside one of the chart templates. i hate that this is even possible and i want to die.
|
# ? Jul 14, 2021 20:14 |
|
jenkins-job-builder supports yaml macros across files, so you can define all your macros in one, separate file. Of course you could that up with common macros in a third file, and then dry that up further in a fourth- and this idea is so good it needs to be an important feature requests for the next major version of kubernetes. hell, maybe for yaml in general.
|
# ? Jul 14, 2021 20:42 |
|
Ansible Adams posted:ngl, what I'm starting to like even more than YAML is HCL2 hcl2 is p. good for what its for
|
# ? Jul 14, 2021 20:52 |
|
but it is explicitly a configuration format, nothing else
|
# ? Jul 14, 2021 20:52 |
|
CMYK BLYAT! posted:yah as someone who maintains a chart i can affirm that Helm is in many ways quite poo poo. really wish they had a proper kustomize integration so all the many many feature requests i get for "add this new deployment feature into the deployment" could be handled with "this doesn't need a loop or advanced template functionality, it is just stuffing a YAML blob into a defined place in another YAML blob, which is exactly what kustomize does, use that instead" but their solution for that is "uh you can run a bash script that invokes kustomize" which is poo poo. haha oh poo poo, last i worked with helm charts putting even a variable in values.yaml wasn't allowed
|
# ? Jul 14, 2021 22:00 |
|
CMYK BLYAT! posted:yah as someone who maintains a chart i can affirm that Helm is in many ways quite poo poo. really wish they had a proper kustomize integration so all the many many feature requests i get for "add this new deployment feature into the deployment" could be handled with "this doesn't need a loop or advanced template functionality, it is just stuffing a YAML blob into a defined place in another YAML blob, which is exactly what kustomize does, use that instead" but their solution for that is "uh you can run a bash script that invokes kustomize" which is poo poo. We have some helm templates that are more brackets than text, and seriously abuse helm templating. All so someone can define all resources they want directly in the values.yaml. It could be worse, another team is using terraform for k8s deployment and every deployment diff is 3+ screens of scrolling.
|
# ? Jul 15, 2021 00:48 |
|
i use helm templates but only for two things: - generating an arbitrary random token for a Secret if the Secret doesn't already exist, for things like internal creds between services - embedding config files into configmaps without needing to have a nested copy in the configmap yaml and i never use other people's helm charts because they're all the union of every possible option that every passerby needed so that it would run on their SPARC cluster with 7 bit bytes or whatever. if a helm chart is the only option then ill render it to yaml with default settings and then edit/use the result because otherwise its hard to tell wtf the thing is doing when its 89% templating logic
|
# ? Jul 15, 2021 01:27 |
|
https://github.com/mittwald/kubernetes-secret-generator i mean this thing exists, although "oops sorry about that time we hosed up the security of the randomly generated values" admittedly doesn't inspire the greatest amount of confidence. helm is 100% terrible
|
# ? Jul 15, 2021 04:16 |
|
rotor posted:so just a general comment here, if you're at work and someone says they're writing their own configuration file format you should respond by setting up a 1-on-1 meeting with them and breaking all their fingers. my favourite hosed up custom config file format is digital anvil's variant of INI for populating databases of object properties where [section]s are actually a class type INI code:
|
# ? Jul 15, 2021 07:38 |
|
Its time to be honest with you all I don’t know what kuberneetus is
|
# ? Jul 15, 2021 19:14 |
|
Cerberus911 posted:It could be worse, another team is using terraform for k8s deployment and every deployment diff is 3+ screens of scrolling. what the hell are they doing, don't they know they can break out configurations into smaller files for specific aspects of services if they're complex?
|
# ? Jul 15, 2021 19:16 |
|
echinopsis posted:Its time to be honest with you all i went to several kubernetes meetups for free food and im still not sure
|
# ? Jul 15, 2021 19:17 |
|
Perplx posted:i went to several kubernetes meetups for free food and im still not sure was the food ok? I went to a PR event last night for lyre non alcoholic spirits. I ate a mushroom canapé that almost made me gag. then some older woman ranted on to me about how I need to change my life. she was kinda hot but she was quite annoying.
|
# ? Jul 15, 2021 19:37 |
|
yes it was good, free pizza, beer, fancy cupcakes, fancy sandwiches, then they stopped serving beer and it wasn't fun anymore
|
# ? Jul 15, 2021 19:59 |
|
Progressive JPEG posted:i never use other people's helm charts because they're all the union of every possible option that every passerby needed there's some very telling bit in a Helm commentary "why this feature is the way it is" post to the effect of "when we started this project, the idea was that you'd maintain your own chart for filling out the specifics of your environments, prod, test, etc. the author and user were expected to be the same person. in practice, we have wound up with chef recipes, where the users are using charts as a shortcut to not understanding how the app they want is deployed" helm would probably be a much saner place if general user audience charts weren't really a thing
|
# ? Jul 15, 2021 20:28 |
|
|
# ? May 3, 2024 05:45 |
|
i work on middleware/developer tools so i only have access to little 3controller/3worker clusters, how big is a decent production kube cluster, anyways?
|
# ? Jul 15, 2021 22:20 |