Hi guys, There was a problem with the

terraform-snowflake-provider

and they fixed it and updated to

2.7.0

pulumi-snowflake

updated from

2.5.0

->

2.6.0

I am trying to update my

pulumi-snowflake

version to the latest one which is

2.6.0

. Inside that update there was a PR that was merged to update the

terraform-snowflake-provider

to

2.7.0

my

requirements.txt

points to

pulumi-snowflake==2.6.0

I cleared the pip cache, removed the plugin from pulumi but when I do pulumi up, it forces it to install

2.5.0

: would anyone know why?

If there's old resources in the state file that are using 2.5.0 then it will be downloaded incase its needed to issue Deletes against. Might end up not being used, but we can't tell when pre-caching at startup.

Hi @echoing-dinner-19531 sorry what do you mean? do I need to delete something ? purge something? or purge my stack?

in other words, if you last ran an

up

with 2.5.0 and you have a resource in your stack (and thus state), then you upgrade to 2.7.0, the program will need both 2.5.0 and 2.7.0 downloaded and present. you may have simultaneously updated the provider to 2.7.0 as well as deleted resources created by 2.5.0. the CLI can't tell, so it precaches 2.5.0 just in case those affected resources need to be deleted. ideal outcome is you should not have to do anything here aside from have a successful

up

. does the

preview

show a concerning diff?

so if it just downloads it, its ok, but then I still get the issue that the fix was supposed to resolve. I updated

pulumi-snowflake

to

2.6.0

which should be pointing to the latest

terraform-snowflake-provider

here you can see the snowflake provider is

2.5.0

but we can see the terraform snowflake provider is at 2.7.0

ah, I'm gonna let @echoing-dinner-19531 take that one when they have a minute 😄

also FYI : thats the output of

pip freeze | grep pulumi

:

pulumi==3.197.0
pulumi_aws==7.6.0
pulumi_command==1.1.0
pulumi_snowflake==2.6.0

You've covered over the plan but I'm assuming that's a create or update?

its a

pulumi up

for a create, just did the command,

RegsiterResource

not found in the logs

Hello, I am going to reply to this thread here. @gray-art-92380 asked me via DM if we're sure this provider includes the lates updates to upstream. It does: • latest release upstream: https://github.com/snowflakedb/terraform-provider-snowflake/releases/tag/v2.7.0 (note the git hash) •

upstream

is updated to point to that hash: https://github.com/pulumi/pulumi-snowflake/pull/976/files#diff-1581e27de87bffae0bd4d745cd7964e68528d7a83e2e4c259a782d275df6f558 • You wouldn't see this in the provider's go.mod because it's there via

replace

.

exactly, thats what I see in the code, its using upstream

Reading this thread, I don't understand the problem entirely. It sounds like you're using both versions somehow?

when doing pulumi up, it downloads the

2.5.0

plugin

You should definitely be seeing a

RegisterResource

call in your logs what happens if you run

pulumi install

? and are there any explicit providers with pinned versions in your state?

so pulumi install seemed to have resolved the issue :

huzzah! 🎉

thats why :

That's not quite why. You did clear out the plugin cache, unless you had something stuck in your PATH.

I did 😓

And "the registry" has nothing to do with it. I'm just cautioning you to take the AI explanation with a grain of salt. I'm glad we got it straightened out - we should definitely make sure that upgrades don't turn into confusion or accidentally pointing to the wrong plugins.

Yeh that AI explanation sounds like rubbish, the cache is keyed by version

ya sorry just trying to understand thats why I tried with AI to understand why @shy-arm-32391 solution worked. I did

pip install

here also did the logtostrderr too here