TimestreamInfluxDB DB instance add more updatable arguments #7
Conversation
…into f-timestreaminfluxdb_db_instance-more-updatable-arguments
Community NoteVoting for Prioritization
For Submitters
|
github-actions
bot
added
documentation
|
After this PR is reviewed, I will close it and create an upstream PR. |
| @@ -494,6 +519,10 @@ func (r *resourceDBInstance) Update(ctx context.Context, req resource.UpdateRequ | |||
| plan.SecondaryAvailabilityZone = flex.StringToFrameworkLegacy(ctx, output.SecondaryAvailabilityZone) | |||
| } | |||
|
|
|||
| if plan.SecondaryAvailabilityZone.IsUnknown() { | |||
| plan.SecondaryAvailabilityZone = state.SecondaryAvailabilityZone | |||
There was a problem hiding this comment.
This is the fix for secondary_availability_zone being left as unknown when tags are updated. If the new value (in plan) is unknown, set it to its old value (in state).
From what I can determine, the problem is that Update is called when tags are updated (for some reason) but UpdateDbInstance isn't actually called, as UpdateDbInstance doesn't allow tags to be updated, therefore secondary_availability_zone is left as unknown. I would like a better fix that scaled better for possible future computed values, but at this point I'm not sure what that better fix would be.
There was a problem hiding this comment.
might be good to add a small comment regarding this in the code 👍
| @@ -494,6 +519,10 @@ func (r *resourceDBInstance) Update(ctx context.Context, req resource.UpdateRequ | |||
| plan.SecondaryAvailabilityZone = flex.StringToFrameworkLegacy(ctx, output.SecondaryAvailabilityZone) | |||
| } | |||
|
|
|||
| if plan.SecondaryAvailabilityZone.IsUnknown() { | |||
| plan.SecondaryAvailabilityZone = state.SecondaryAvailabilityZone | |||
There was a problem hiding this comment.
might be good to add a small comment regarding this in the code 👍
| }, | ||
| Validators: []validator.Int64{ | ||
| int64validator.Between(1024, 65535), | ||
| int64validator.NoneOf(2375, 2376, 7788, 7789, 7790, 7791, 7792, 7793, 7794, 7795, 7796, 7797, 7798, 7799, 8090, 51678, 51679, 51680), |
There was a problem hiding this comment.
(jk)
| int64validator.NoneOf(2375, 2376, 7788, 7789, 7790, 7791, 7792, 7793, 7794, 7795, 7796, 7797, 7798, 7799, 8090, 51678, 51679, 51680), | |
| int64validator.NoneOf(1337, 2375, 2376, 7788, 7789, 7790, 7791, 7792, 7793, 7794, 7795, 7796, 7797, 7798, 7799, 8090, 51678, 51679, 51680), |
There was a problem hiding this comment.
Is that a leetspeak reference?
| @@ -215,6 +211,18 @@ func (r *resourceDBInstance) Schema(ctx context.Context, req resource.SchemaRequ | |||
| also use the InfluxDB CLI to create an operator token. These attributes will be | |||
| stored in a Secret created in AWS SecretManager in your account.`, | |||
| }, | |||
| names.AttrPort: schema.Int64Attribute{ | |||
| Optional: true, | |||
| Computed: true, | |||
There was a problem hiding this comment.
would it be redundant to add default here as well?
| Computed: true, | |
| Computed: true, | |
| Default: int64(8086), |
There was a problem hiding this comment.
It would be redundant, the port defaults to 8086 anyways when creating an instance.
Add computed to service region
Description
network_typeadded as an argument.portadded as an argument.deployment_typeis now updatable.TestAccTimestreamInfluxDBDBInstance_deploymentType.db_instance_typeis now updatable.TestAccTimestreamInfluxDBDBInstance_dbInstanceTypehas been added with an update step.flex.Expandis no longer used inUpdateto populate anUpdateDbInstanceInputstruct. This is because aValidationExceptionwill occur when a DB instance is updated with configuration values it already uses.Update, ifsecondary_availability_zoneis left as unknown, it is set to its previous value.secondary_availability_zonecan be left as unknown when tags are updated, which causes an error.secondary_availability_zoneuses the default value ofnil. While unconventional, this prevents an error when updatingdeployment_type.secondary_availability_zone's value is directly a result of the value ofdeployment_type.awstypes.StatusDeletedhas been added as a pending state inwaitDBInstanceDeleted. This is because DB instances are first given the status"DELETED"before becoming unreachable. This status is not consistent. Using this status as a target status leads to flaky tests. Using this status as a pending status means Terraform will wait until the instance is unreachable.Relations
N/A.
References
Updated
UpdateDbInstanceInputstruct documentation.Output from Acceptance Testing