https://pulumi.com logo
Title
r

rapid-appointment-67477

03/28/2023, 10:18 PM
Hello. I'm trying to major version upgrade a Postgres RDS resources in AWS (13.8 to 14.6) and am seeing an error. I haven't been finding much on the error via google, but I did find this slack channel so I thought I'd ask here! I changed this
aws.rds.Cluster(engine_version="14.6")
and am getting this message
Failed to modify RDS Cluster (cluster-name) InvalidParameterCombination: The current DB cluster parameter group parameter-group-name is custom. You must explicitly specify a new DB cluster parameter group, either default or custom, for the engine version upgrade.
the error suggested creating a new cluster parameter group, so I changed this original code:
pg_name = "parameter-group-name"
parameter_group = ClusterParameterGroup(
    pg_name,
    name=pg_name,...)

Cluster(...
    db_cluster_parameter_group_name=parameter_group.name,...)
I only changed
pg_name="parameter_group-name-versioned"
and now pulumi shows it wants to create the new parameter group but still get the same error. what else is needed?
s

steep-toddler-94095

03/28/2023, 11:02 PM
what are the params in
ClusterParameterGroup
? I would guess it's still referncing the old DB version
r

rapid-appointment-67477

03/28/2023, 11:07 PM
Other than the family (I updated that
14
from
13
, I don't see anything version specific.. unless you mean one of the parameters is no longer supported?
ClusterParameterGroup(
        parameter_group_name,
        name=parameter_group_name,
	family="aurora-postgresql14",
        parameters=[
            ClusterParameterGroupParameterArgs(
                name="rds.logical_replication",...
            ),
            ClusterParameterGroupParameterArgs(
                name="rds.force_ssl",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_min_duration_statement",...
            ),
            ClusterParameterGroupParameterArgs(
                name="shared_preload_libraries",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_destination",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_filename",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_hostname",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_rotation_age",...
            ),
            ClusterParameterGroupParameterArgs(
                name="log_statement",...
            ),
            ClusterParameterGroupParameterArgs(
                name="rds.log_retention_period",...
            ),
            ClusterParameterGroupParameterArgs(
                name="max_parallel_workers",...
            ),
            ClusterParameterGroupParameterArgs(
                name="max_parallel_workers_per_gather",...
            ),
            ClusterParameterGroupParameterArgs(
                name="max_parallel_maintenance_workers",...
            ),
        ],
    )
s

steep-toddler-94095

03/29/2023, 1:54 AM
No, I did mean the family. Hmm i am not sure what's causing the issue then. Is it possible for you to work around this by changing the current parameter group to the default one before attempting the upgrade? Or do you really depend on some of the custom params you set
There's this issue related to cloudformation but I'm wondering if you are also seeing the same issue here: https://github.com/aws-cloudformation/cloudformation-coverage-roadmap/issues/1373
r

rapid-appointment-67477

03/29/2023, 2:24 AM
tomorrow I'll try the steps in the issue 1373 link provided