fix: Add workaround to ignore groups config when provisioning instance using backup CRN (#293)

Author: jor2

main.tf

@@ -27,6 +27,9 @@ locals {
   # Determine if host_flavor is used
   host_flavor_set = var.member_host_flavor != null ? true : false
 
+  # Determine if restore, from backup or point in time recovery
+  recovery_mode = var.backup_crn != null || var.pitr_id != null
+
   # Determine what KMS service is being used for database encryption
   kms_service = var.kms_key_crn != null ? (
     can(regex(".*kms.*", var.kms_key_crn)) ? "kms" : (
@@ -84,11 +87,15 @@ resource "ibm_database" "enterprise_db" {
     }
   }
 
+  # Workaround for https://github.ibm.com/GoldenEye/issues/issues/11359
+  # means that no `group` block is added when restoring from backup
+  # or point in time recovery
+
   ## This for_each block is NOT a loop to attach to multiple group blocks.
   ## This is used to conditionally add one, OR, the other group block depending on var.local.host_flavor_set
-  ## This block is for if host_flavor IS set to specific pre-defined host sizes and not set to "multitenant"
+  ## This block is for if host_flavor IS set to specific pre-defined host sizes
   dynamic "group" {
-    for_each = local.host_flavor_set ? [1] : []
+    for_each = local.host_flavor_set && !local.recovery_mode ? [1] : []
     content {
       group_id = "member" # Only member type is allowed for IBM Cloud Databases
       host_flavor {
@@ -112,7 +119,7 @@ resource "ibm_database" "enterprise_db" {
 
   ## This block is for if host_flavor IS NOT set
   dynamic "group" {
-    for_each = local.host_flavor_set ? [] : [1]
+    for_each = !local.host_flavor_set && !local.recovery_mode ? [1] : []
     content {
       group_id = "member" # Only member type is allowed for IBM Cloud Databases
       memory {
@@ -172,6 +179,8 @@ resource "ibm_database" "enterprise_db" {
 
   timeouts {
     create = "120m" # Extending provisioning time to 120 minutes
+    update = "120m"
+    delete = "15m"
   }
 }