oci_database_cloud_vm_cluster on too many db_servers

[ad-mohammad]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform Version and Provider Version

Terraform v1.14.5
on linux_amd64

• provider registry.terraform.io/hashicorp/azurerm v4.46.0
• provider registry.terraform.io/oracle/oci v7.13.0

Affected Resource(s)

oci_database_cloud_vm_cluster

Terraform Configuration Files

# Copy-paste your Terraform configurations here - for large Terraform configs,
# please use a service like Dropbox and share a link to the ZIP file. 
# Please remove any sensitive information from configuration files before sharing them. 

Debug Output

Panic Output

Expected Behavior

I have an Infra with 4 dbServers. A vmcluster is created on this Infra with parameter db_servers referencing dbServer3 and dbServer4 only. The cluster should be creating on only these two dbServers and should then persist on only these two.

Actual Behavior

Initial creation of the vmcluster succeeds and the cluster is created on only dbServers3 and dbServer4. With no further configuration changes I run a plan again and the module want to modify the cluster db_servers. To see the effect I then run an apply and the cluster is expanded to be on all four available dbServers despite the parameter db_servers being unchanged in the config file.
Terraform will perform the following actions:

module.vmcluster_vmc_prd_retailX11.oci_database_cloud_vm_cluster.cluster will be updated in-place
 ~ resource "oci_database_cloud_vm_cluster" "cluster" {
     ~ db_servers                             = [
         - "ocid1.dbserver.oc1.eu-frankfurt-1.---4pauuwwceimrcaqmfgcapkaxa",
           "ocid1.dbserver.oc1.eu-frankfurt-1.---sqnm7jgk72mdx4acx6inbbmza",
         + "ocid1.dbserver.oc1.eu-frankfurt-1.---4pauuwwceimrcaqmfgcapkaxa",
       ]
       id                                     = "ocid1.cloudvmcluster.oc1.eu-frankfurt-1.---zpwvpyqtmbw4nvi62q2uhxqza"
       # (53 unchanged attributes hidden)

       # (14 unchanged blocks hidden)
   }

Plan: 0 to add, 1 to change, 0 to destroy.

Steps to Reproduce

create 2 node vmcluster on 4 node Infra then run plan. Change indicated where no parameter has been modified since the create.

1. terraform apply

Important Factoids

I can reproduce this in our two regions AMS and FRA

References

[tf-oci-pub]

Thank you for reporting the issue. We have raised an internal ticket to track this. Our service engineers will get back to you.

[acui107]

Please help provide the TF config files including the full OCIDs for us to review the history

[ad-mohammad]

ExaInfra = ocid1.cloudexadatainfrastructure.oc1.eu-frankfurt-1.antheljsnyqczmqaqbmon5z7l7gbsm4ksrolwb4yvnesq7imwq26rje7fv5q
VMcluster = ocid1.cloudvmcluster.oc1.eu-frankfurt-1.antheljsnyqczmqapsk75leu6n5edxrc7vhzpwvpyqtmbw4nvi62q2uhxqza

[ad-mohammad]

The is the relevant config file we use to create the vmcluster where we hardcode the ocid's for the db_servers:

module "vmcluster_vmc_prd_retailX11" {
  source         = "../../../../modules2/oci/db-exadata-vmclusterX11"
  providers = {
    oci = oci.frankfurt
  }
  display_name   = module.naming_vmcluster_vmc_prd_retailX11.cloud_exadata_vmcluster.name
  compartment_id = data.oci_identity_compartments.destination_cmp.compartments[0].id
  infra_id       = module.servers_infra_retailX11.infrastructure_id
  networking = {
    db_subnet_id     = local.subnet_client_retail_id
    backup_subnet_id = local.subnet_backup_id
    db_subnet_nsg_ids     = local.retail_client_nsg_list
    backup_subnet_nsg_ids = local.backup_nsg
  }
  cluster_name = "fr04prdrtl" 
  hostname     = "fr04prdrtl" 
  shape = {
    cpu_core_count      = 32 # Minimum is 16 when creating new using X11M - this value is the total for the cluster
    memory_gb       = 300
    node_storage_gb = 600
  }
  data_storage = {
    storage_tbs        = 100
    storage_percentage = 80
  }
  db_servers   = ["ocid1.dbserver.oc1.eu-frankfurt-1.antheljsc7zwxhyazq7yrwa5jnvb4kdr3ndsqnm7jgk72mdx4acx6inbbmza","ocid1.dbserver.oc1.eu-frankfurt-1.antheljsc7zwxhyaieoimkabhxraywllmnh4pauuwwceimrcaqmfgcapkaxa"]
  defined_tags = local.defined_tags

  # Immutable properties
  gi_version                  = "23.0.0.0"
  initial_ssh_public_key      = data.azurerm_key_vault_secret.vmc_prd_retailX11.value
  timezone                    = "CET"
  is_local_backup_enabled     = false
  is_sparse_diskgroup_enabled = false
  scan_listener_port_tcp      = 1521
  scan_listener_port_tcp_ssl  = 2484
}

The source module runs this:

resource "oci_database_cloud_vm_cluster" "cluster" {
  # Placement
  display_name                    = var.display_name
  cloud_exadata_infrastructure_id = var.infra_id
  compartment_id                  = var.compartment_id

  # Networking
  subnet_id              = var.networking.db_subnet_id
  nsg_ids                = var.networking.db_subnet_nsg_ids
  backup_subnet_id       = var.networking.backup_subnet_id
  backup_network_nsg_ids = var.networking.backup_subnet_nsg_ids
  cluster_name           = var.cluster_name
  hostname               = var.hostname

  # Initial cluster properties (not updatable)
  gi_version                  = var.gi_version
  ssh_public_keys             = [var.initial_ssh_public_key]
  time_zone                   = var.timezone
  is_local_backup_enabled     = var.is_local_backup_enabled
  is_sparse_diskgroup_enabled = var.is_sparse_diskgroup_enabled
  scan_listener_port_tcp      = var.scan_listener_port_tcp
  scan_listener_port_tcp_ssl  = var.scan_listener_port_tcp_ssl

  # Cluster shaping
  cpu_core_count              = var.shape.cpu_core_count
  memory_size_in_gbs          = var.shape.memory_gb
  db_node_storage_size_in_gbs = var.shape.node_storage_gb
  data_storage_size_in_tbs    = var.data_storage.storage_tbs
  data_storage_percentage     = var.data_storage.storage_percentage
  db_servers                  = var.db_servers

  # Tagging
  defined_tags  = var.defined_tags
  freeform_tags = var.freeform_tags

  # Others
  license_model = var.license_model
  data_collection_options {
    is_diagnostics_events_enabled = var.data_collection_options.is_diagnostics_events_enabled
    is_health_monitoring_enabled  = var.data_collection_options.is_health_monitoring_enabled
    is_incident_logs_enabled      = var.data_collection_options.is_incident_logs_enabled
  }

  lifecycle {
    ignore_changes = [
      ssh_public_keys,             # Ignored as keys will be rotated externally to the module
      gi_version,                  # Ignored as this will change during upgrades
      time_zone,                   # Not updatable
      is_local_backup_enabled,     # Not updatable
      is_sparse_diskgroup_enabled, # Not updatable
      create_async
    ]
  }
}

[ad-mohammad]

I have tried updating to a more recent release of the modules but the problem is the same when using:
Terraform v1.14.5
on linux_amd64

• provider registry.terraform.io/hashicorp/azurerm v4.59.0
• provider registry.terraform.io/oracle/oci v8.0.0