Flag to wait till deployment actually succeeded

[yellowmegaman]

Hello there! Question here.

It would be really great if i could run nomad run and it won't release terminal till deployment is not only scheduled, but also completed successfully.

Use case:

• all deployments are being run from CI
• template nomad job against some environment via consul-template
• or template nomad job via terraform (template_file)
• plan/run template to deploy stuff
• terminal gets released if constrains/capacity are OK
• run script to check whenever deployment is successful by getting desiredCount and comparing to healthyCount in a loop

This approach is sub-optimal since it relies on some scripting which can be faulty and may have some corner cases.

Can we add flag to nomad run so it won't throw 0 exit code till deployment is finished?
Same would be awesome for nomad provider for terraform.

Thanks for attention in advance!

[jippi]

a work-around for now is to use jrasell/levant

[yellowmegaman]

@jippi thanks! seen it before, never had a chance to use it yet.

In my opinion it's not something that should be external to nomad, since it is already aware of the state of deployed services.

Anyway, gonna try levant for now, leaving issue open in case there would be some update whenever feature may land on roadmap for some future release.

[jippi]

Yes, I'm pretty sure Nomad will eventually support this - should be pretty trivial to add - just wanted to unblock you until Nomad got native support :)

[yellowmegaman]

@jippi in that you've succeeded =) Thank again

[yellowmegaman]

Well, levant seems to be completely broken, deploying nomad job with broken on purpose healthcheck results in success, so hope this issue will see some attention someday.

[jrasell]

@yellowmegaman feel free to raise an issue against Levant and I can take a look into it. We use it extensively and successfully in my current job so I would be curious what the issue is.

[dadgar]

@yellowmegaman I do think it is a good feature and would want to add it. @jrasell PRs welcome 😄

[yellowmegaman]

@dadgar that is super-awesome =)
@jippi hashicorp/levant#227

[cgbaker]

@yellowmegaman : FYI, i'm adding this feature to the Nomad TF provider:
hashicorp/terraform-provider-nomad#56

[pznamensky]

Looks like levant is abandoned by owner. Thus the flag in Nomad would be super useful for us too.

[pznamensky]

Is there any news here? :)
Unfortunately it blocks upgrade to nomad 1.0 for us.
We used to watch for deployments using levant, but it doesn't have hcl2 support (hashicorp/levant#400)

[tcurdt]

No hcl2 support is a blocker to use levant.
And I also agree this should be ideally be built into nomad.

A nomad run should have an interactive option. Either interactive by default with a background/daemon option (see docker-compose -d) or non-interactive by default and allowing to activate interactive (see docker -i).

Not seeing allocation and deployments details as well as process stdout/err on a nomad run is a terrible UX.
Worse - there is not even a simple way to get to this information like that from the command line.
Why is there no nomad alloc list for example?

[thatsk]

we want to see this feature in nomad cli. So that we can make ci-cd via nomad cli to get realtime deployment in logs

[tgross]

This issue is a duplicate of #6818, which is on the roadmap. I'm going to close this issue in lieu of #6818 only because it has more of the implementation details on how we want to proceed on this. Keep an eye there for future progress (also: our Product folks do look at emoji reactions to issues, so feel free to add them there!)

[github-actions]

I'm going to lock this issue because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.