could you please prioritize <https://github.com/pu...
# general
g
could you please prioritize https://github.com/pulumi/pulumi-gcp/issues/55 ? This is kind of blocking me from converting a lot of terraform code to pulumi.
s
Yes - can do! I’m not so familiar with GCP - is a NodePool something that stands on its own or are there other pieces around it?
One thing I can tell you is that the proximate cause of the panic is that it expects
strings.split(id, "/")[2]
to be a thing - which it is not given the name there.
Is there a more qualified version of the name that might be required?
g
ah i see
yeah maybe
a nodepool belongs to a cluster
so maybe i have to qualify it with the clustername
going to try that out…
ok I checked a bit what terraform itself expects as identifier when doing
terraform
import
turns out it expects
<zone>/<cluster_name>/<pool_name>
and when using this in pulumi as second param it actually works.
So I guess the real problem here is that pulumi errors out with some useless error message (whereas terraform gives an actually useful error message). Let me grab the error message that terraform would give when using an incorrect identifier…
ok terraforms error message looks like this when attempting to use simply
default-pool
as identifier:
Copy code
google_container_node_pool.default_pool: Importing from ID "default-pool"...

Error: google_container_node_pool.default_pool (import id: default-pool): import google_container_node_pool.default_pool (id: default-pool): Invalid container cluster specifier. Expecting {zone}/{cluster}/{name} or {project}/{zone}/{cluster}/{name}
Actually pretty helpful 🙂
w
This is ultimately because of https://github.com/pulumi/pulumi-terraform/issues/266. The TF error above is generated by the call to Import. We don’t currently call that, but Read will not work without that Importer having run.
s
@glamorous-printer-66548 I think we have a way forward to fix this issue - I’ll likely try it out tomorrow now.
👍 1