archived the channel

@cflowe has joined the channel

@Rizky Ramadhan has joined the channel

@william has joined the channel

@Karbust has joined the channel

@Mohammed Yahya has joined the channel

Adding @U010XGY9B46 bot

@UUB28NLDS help keep tabs! 😉

@Tertius Stander has joined the channel

set the channel description: https://github.com/joshmyers/dynolocker
Archive: https://archive.sweetops.com/dynolocker/

@richwine has joined the channel

huh, I like this Variant 😄

Yeah, I was going to try and do more with it and actually have it do sane things with process and signals etc but never got around to it

(And already use it for terraform locking)

But yea, I get your point. For our purposes, AWS is an acceptable requirement.

Haha well, kinda depends on what “outside” means ;-)

well, def won’t 🙂

dynolocker might not work well outside of AWS

That might not work well if running outside of k8s

Yeah it was based off etcd

That one requires etcd, no?

locksmithctl by coreOS may also be interesting

it was a hacky tool whipped up due to a race condition when bootstrapping Vault nodes (a few years back now)

Sure, PRs always welcome

Morning 👋

not yet, but i’m seriously considering to move to codefresh for anything related to github flow

are you using codefresh with terraform now?

i think for terraform it's scarier

but i do have encountered the same situation as erik’s in my terraform pipeline powered by github flow(the latest one, apply/deploy before merging for final testing)

first release then was blocked by second release completing

(happened to me yesterday... and we have wait: true, so second release was rolling out destroying pods from first release)

probably it is unlikely to happen when you run the pipeline on a new commit to e.g. master

set the channel topic: https://github.com/joshmyers/dynolocker

that's (2) very strong use-cases for locking

we should really be locking a helm release before performing a helm deployment

and we end up doing concurrent helm deployments of the same app

also, i've had the problem we're we merge 2 PRs too close to each other

(e.g. like in atlantis)

for example, if you're terraforming, you want to lock a project to a pull request

basically, you need to lock something which could be used across pipelines or even within the same pipeline

see the snippet above

so the use case is...?

then we can also distribute through cloudposse/packages

also wire it up with a codefresh pipeline to build binary releases

yea, let's breath some life into it

i definitely like a command written in golang 🙂 am eager to submit a pr to dynolocker if that makes sense

ah i remembered that i had created a very similar command myself 3 years ago, but in ruby
https://github.com/crowdworks/joumae-ruby#cli

I was also sharing dynolocking with @dustinvb from codefresh

I don't remember