Classic VM is only really deleted some time after terraform destroy complete

[l2fprod]

Using this sample https://github.com/IBM-Cloud/LAMP-terraform-ibm, I did:

• configure my variables,
• terraform init,
• terraform apply.

Then I did terraform destroy

Everything got destroyed but for a brief moment I could still connect to the VSI after terraform completed until it got really deleted:

I could also see the device in the UI console.

I would expect terraform returns only after the resources have been confirmed to be really deleted/unreachable

[stevestrutt]

@l2fprod that is the way that delete of a Classic VSI works. The delete request is async and IMS then cleans up the VSI's using a cron job run on a 20 minute schedule. This is the behaviour of the SL API. Hoping for better with the VPC apis.

[l2fprod]

Maybe this should be part of the provider documentation. I was really surprised I could still connect and thought the delete did not work.

Is it the same in VPC? Delete is also async there

[hkantare]

In VPC its not an async..The vm gets deleted when user run destroy.

[l2fprod]

@hkantare @davetropeano

you mean that for VPC the terraform provider is waiting for the VSI to actually disappear? As the delete process of a VSI in VPC is async too - VSI first goes to "Deleting" state before disappearing. Then why not doing the same for the classic VSI?

[hkantare]

@geethasathyamurthy can you please update this in compute VM instance documentation