[Redis]: Updating redis instance with parameters (memory and notify_keyspace_events) in a single request fails

[gurubanigulati17]

Is there an existing issue for this?

• I have searched the existing issues

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 and review the contribution guide to help.

Terraform Version

0.15.5

AzureRM Provider Version

3.96.0

Affected Resource(s)/Data Source(s)

azurerm_redis_cache

Terraform Configuration Files

# redis.tf.json

{ 
   "resource": {
        "azurerm_redis_cache": {
            "redis_cache": {
                "name": "redis-42894ed2-15d8-44d2-8e74-a5602ad17a11",
                "location": "westeurope",
                "resource_group_name": "azure-staging-backing_services_infra",
                "redis_version": "6",
                "capacity": 3,
                "family": "P",
                "sku_name": "Premium",
                "enable_non_ssl_port": false,
                "minimum_tls_version": "1.2",
                "redis_configuration": {
                    "maxmemory_policy": "allkeys-lru",
                    "notify_keyspace_events": "AKE"
                },
                "patch_schedule": {
                    "day_of_week": "Saturday",
                    "start_hour_utc": 22,
                    "maintenance_window": "PT5H"
                },
                "shard_count": 2,
                "subnet_id": "xxx",
                "zones": [
                    "1",
                    "2",
                    "3"
                ],
                "timeouts": {
                    "create": "8h",
                    "update": "8h",
                    "delete": "8h"
                },
                "lifecycle": {
                    "prevent_destroy": true
                }
            }
        }
}

# terraform.tfstate

 {
      "mode": "managed",
      "type": "azurerm_redis_cache",
      "name": "redis_cache",
      "provider": "provider[\"registry.terraform.io/hashicorp/azurerm\"]",
      "instances": [
        {
          "schema_version": 1,
          "attributes": {
            "capacity": 3,
            "enable_non_ssl_port": false,
            "family": "P",
            "redis_configuration": [
              {
                "active_directory_authentication_enabled": false,
                "aof_backup_enabled": false,
                "aof_storage_connection_string_0": "",
                "aof_storage_connection_string_1": "",
                "data_persistence_authentication_method": "SAS",
                "enable_authentication": true,
                "maxclients": 15000,
                "maxfragmentationmemory_reserved": 1330,
                "maxmemory_delta": 1330,
                "maxmemory_policy": "allkeys-lru",
                "maxmemory_reserved": 1330,
                "notify_keyspace_events": "AKE",
                "rdb_backup_enabled": false,
                "rdb_backup_frequency": 0,
                "rdb_backup_max_snapshot_count": 0,
                "rdb_storage_connection_string": "",
                "storage_account_subscription_id": ""
              }
            ],
            "redis_version": "6.0" }]
}

Debug Output/Panic Output

[ERROR] provider.terraform-provider-azurerm_v3.96.0_x5: Response contains error diagnostic: diagnostic_severity=ERROR diagnostic_summary="updating Redi (Subscription: "xxx"
Resource Group Name: "azure-staging-backing_services_infra"
Redis Name: "redis-42894ed2-15d8-44d2-8e74-a5602ad17a11"): unexpected status 400 with error: BadRequest: The following updates can't be processed in one single request, please send separate request to update them: 'properties.sku.capacity,properties.aadEnableDisable'.
RequestID=f2a51a49-38b1-40e3-bda0-8cc8dccd25b9" @caller=github.com/hashicorp/[email protected]/tfprotov5/internal/diag/diagnostics.go:58 @module=sdk.proto diagnostic_detail= tf_proto_version=5.4 tf_provider_addr=provider tf_req_id=ceb6c5b7-c0f2-811b-c8a9-c65db4f4a983 tf_resource_type=azurerm_redis_cache tf_rpc=ApplyResourceChange timestamp=2024-05-19T18:46:09.807ZError: updating Redi (Subscription: "xxx"
Resource Group Name: "azure-staging-backing_services_infra"
Redis Name: "redis-42894ed2-15d8-44d2-8e74-a5602ad17a11"): unexpected status 400 with error: BadRequest: The following updates can't be processed in one single request, please send separate request to update them: 'properties.sku.capacity,properties.aadEnableDisable'.
RequestID=f2a51a49-38b1-40e3-bda0-8cc8dccd25b9  with azurerm_redis_cache.redis_cache,
  on redis.tf.json line 67, in resource.azurerm_redis_cache.redis_cache:

Expected Behaviour

The operation ought to proceed without the error message indicating, "The following updates can't be processed in one single request, please send separate request to update them: 'properties.sku.capacity,properties.aadEnableDisable". In versions 3.82.0 and below, this operation is supported. However, issues arise with versions beyond 3.82.0, wherein the operation fails. In addition, I have conducted tests on all the versions including 3.83.0, 3.84.0, 3.96.0, 3.97.0, 3.98.0, and 3.102.0.

Actual Behaviour

We attempted to trigger an update operation which increases the memory from 16 to 32 GB and changes notify_keyspace_events from AK to AKE in one call. However, it resulted in the failure of this update instance operation.

The error message received was 'The following updates can't be processed in one single request, please send separate request to update them: 'properties.sku.capacity,properties.aadEnableDisable'.
This suggests that these changes should be requested separately.

Steps to Reproduce

1. Setup a new Azure Redis Server with

{
  "cluster_mode": true,
  "engine_version": "6.0",
  "eviction_policy": "allkeys-lru",
  "memory": 16,
  "multi_az": true,
  "notify_keyspace_events": "AK",
  "shard_count": 2
}

2. Change the memory to 32 and notify_keyspace_events to AKE.

Important Factoids

No response

References

No response

[thoHeinze]

@favoretti Sorry for pinging you directly, but I see you worked on two PRs related to enable AAD (which is mentioned in the error message above) in Redis: #23976 and #25006 Can you make any sense of the error described by @gurubanigulati17 ? Is this the new expected behaviour?

[favoretti]

@thoHeinze I see what is happening, but I don't know why. I'd need to run a couple of tests to try and reproduce it, but I'm unfortunately swamped hard at work right now, it will take me at least 2-3 weeks before I catch a breath to be able to do anything about this, sorry..