Terraform-provider-digitalocean: Creating PostgreSQL DBs and users randomly fails

[lukoyanov]

Hi.

There is some flaky issue when creating PostgreSQL databases and users via DO provider. Please see attached minimal proof-of-concept.
Terraform complains with the following message:

Error: Provider produced inconsistent result after apply

When applying changes to digitalocean_database_user.db-user, provider
"registry.terraform.io/-/digitalocean" produced an unexpected new value for
was present, but now absent.

This is a bug in the provider, which should be reported in the provider's own
issue tracker.

Terraform Version

Terraform v0.12.19

• provider.digitalocean v1.12.0

Affected Resource(s)

• digitalocean_database_user.db-user
• digitalocean_database_user.db

Expected Behavior

User and database should be successfully created.

Actual Behavior

I'm no sure, but it seems that user's and database's meta information is not put in a state file. Actual resources is mostly created in a cloud.

Steps to Reproduce

Please unzip attached POC, then
$ export TF_VAR_do_token=your-token
$ export TF_VAR_pg_cluser_name=your_pg_cluster_name
$ ./test.sh

POC:
do_bug.zip

[eddiezane]

@lukoyanov apologies you hit this and thanks for reporting.

I checked with the team and they believe this is due to a caching issue our E2E tests recently surfaced. They're already working on a fix! I'll keep you posted as it progresses.

[ivanov-aleksander]

I have also faced this issue

[eddiezane]

We disabled caching here back on Feb 2nd.

@lukoyanov @ivanov-aleksander @rfleschenberg have any of you ran into this since then?

[lukoyanov]

@eddiezane just tried to run a test create-destroy loop and it seems that bug is still there.
I've got the same error:

Error: Provider produced inconsistent result after apply

When applying changes to digitalocean_database_db.db, provider
"registry.terraform.io/-/digitalocean" produced an unexpected new value for
was present, but now absent.

This is a bug in the provider, which should be reported in the provider's own
issue tracker.

Failed to provision infrastructure, retrying...

Please check out my attached do_bug.zip archive which contains create-drop test script.

[lukoyanov]

I've got an error after 8 successful create-destroy iterations.

[rfleschenberg]

We disabled caching here back on Feb 2nd.

@lukoyanov @ivanov-aleksander @rfleschenberg have any of you ran into this since then?

I didn't, but I didn't have the opportunity to try since then.

[lukoyanov]

Does anyone in DO working on this?
My deployment automation is broken for more then a month and now i'm heading over to some ugly curl workarounds :-(

[eddiezane]

@lukoyanov just an update here.

We've got a fix that should be rolling out this week. I've been running your test script in a loop all morning and it hasn't failed.

[eddiezane]

@lukoyanov the fix should be in production now. Thanks for your patience and please reopen if you still experience this.