Can't deploy on environment in state canceled

Bugs
1

Hello,

When we canceled a deployment it’s not possible to redeploy because environment is in state canceled. For that we have to let the deployment failed to be able to deploy again and we loose time when container can’t start or is in crash loop.

It is possible to enable deployment when environment is in state canceled ?

Thanks a lot

2

Hi @Orkin,

I wasn’t able to reproduce the issue you are describing. I have tried to cancel the deployment of an environment and I can relaunch a deployment right after (once my environment deployment is flagged as “cancelled”)

Can you give us more detail on what you have done exactly?

thanks
Alessandro

3

Hi @a_carrano

I use the Qovery github action to deploy my application. In my workflow I specify all the applications id that I need to deploy. I have 3 applications with the same code base that are deploy in same time (I can’t wait parallel deployments :stuck_out_tongue:). All were in crash loop and 2 of them in failed state because of too many crashed. So I cancel the last one deployment to save some time because it’s in crash loop too but not in failed state.
I push my fix to prevent container to crash and deploy again. And in my github action I have this message : Environment cannot accept deploy yet, state: CANCELED

4

thanks

which version of the github app are you using? (We think that the problem comes from there)

today we are at the 0.14.3 and we had before exactly a typo on the canceled status generating this issue

5

I use this version : Qovery/qovery-action@v0.10 I think I’m late :sweat_smile:

6

ok that should definitively solve your problem.

FYI we have updated our CLI and it can be used directly on github action. You can find the documentation here and here the other examples of actions

On top of being always updated in your pipeline (since the script downloads the client), you will find all the latest updates on the CLI (we will invest more in the CLI integration)

1 Like
7

@a_carrano so it’s not usefull to use Qovery github action anymore ?

8

we are still supporting it but we encourage you to move to the implementation using the CLI

1 Like
9

Hi again @a_carrano, I tried to migrate my CI using the CLI but I’m faced with a problem. My application is configured from a github repository with auto-deploy set to false. On the Qovery Cli using container deploy I think we have to use an application with application source set to Registry instead of git provider. I can’t find how to migrate an existing application to application source container without creating a new one. How can I handle it ?

And other question, I have 400 bad request on Qovery Deploy command from Cli and can’t find why. It is possible to run it with debug mode or in verbose mode to get de real error message ? I tried with organization / project / environment and container name and id

Thanks

10

Hi @Orkin , unfortunately, it’s impossible to migrate your app from a git-based repository to a container one. The reason is that it does not rely on the same configuration under the hood. So you will need to create a new application. Maybe we can thing to provide a migration helper via our product if it happens often - cc @a_carrano

Can you show me the command you try to run?

Romaric.

11

@rophilogene qovery container deploy --organization a2ed680d-ffdf-44a2-905f-24fc2a58a97d --project 6de9b4cc-282f-4d25-8a5b-45a043966a99 --environment 249bdc7d-8b9c-48c6-ae49-8ffddd0dc95b --container cac8e170-ea1d-4064-91e2-40cb370c59dd --tag 614cb30 --watch

I follow this guide

No really need for migration tools just the possibility to export all env variable to import them. This is the only thing that is painfull

12

You have to use names and not IDs (it’s more convenient). If you have space in your resource names - double quote them E.g qovery container deploy -n "App Name"

If you have used a scope like Project or Environment then you don’t need to export env vars and secrets. Otherwise, you’ll need to re-import them.

13

@rophilogene I tried with app name that is “api-container” and have always a 400 bad request and can not figure out which api call throw this error :confused:

14

I forget to say that your commit ID must be the long version and not the short one

15

Same :confused: qovery container deploy --organization "Organization" --project "Api" --environment "development" -n "api-container" --tag 614cb30904a912d94f816cb9693349deba71d7ff --watch

I just changed the organization on the exemple here

16

if it’s a container then it can’t be a commid-id unless you tagged your image with a commit id? :thinking:

17

@rophilogene I can put anything in all values I’m always faced in 400 bad request event when data are invalid. It don’t trigger the deployment state on the cluster

18

I take a look at your configuration - and your container uses a tag name develop at the moment. You must provide a valid tag name with the parameter --tag. BTW, I highly recommend using a unique tag name per image. develop is not a unique tag name, and you can have consistency issues between redeployments.

19

Yes I can change that, I have to put one on service creation interface and though it will be already the same but I think it’s only for initial deployment right ?

20

What do you mean? (Sorry, I’m not sure to understand)