Concurrency error in underlying Terraform provider

[nb-ccd]

I have run into the following issue while revoking grants a couple of times:

Diagnostics:
  pulumi:pulumi:Stack (<stack>):
    error: update failed
 
  postgresql:index:Grant (<user>_table_grant):
    error:   sdk-v2/provider2.go:572: sdk.helper_schema: could not execute revoke query: pq: tuple concurrently updated: provider=postgresql@3.16.0
    error: deleting urn:pulumi:<stack>::<program>::postgresql:index/grant:Grant::<user>_table_grant_table_grant: 1 error occurred:
    	* could not execute revoke query: pq: tuple concurrently updated
 
Resources:
    - 12 deleted

This is while running destroy against a stack which provisions two different users with table and sequence level grants.

The bug seems to be mentioned in the Terraform provider here: cyrilgdn/terraform-provider-postgresql#178

In that issue, they say it is possible work around the issue by setting parallelism=1, and there also seems to be a fix that has been tested and found working but not merged by the maintainer. Raising it here because I want to check in whether we can provide a workaround within Pulumi.

[pgavlin]

In that issue, they say it is possible work around the issue by setting parallelism=1

We do have a CLI flag that controls parallelism--you could try setting that to 1. See -p in https://www.pulumi.com/docs/iac/cli/commands/pulumi_up/

Alternatively, you could use try using the dependsOn resource option to keep the operations from executing concurrently regardless of the overall parallelism.