Skip to content

[bugfix] aws_elasticache_replication_group : fixed num_node_groups downscaling #45893

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
YakDriver merged 9 commits into from
Feb 2, 2026
Merged

[bugfix] aws_elasticache_replication_group : fixed num_node_groups downscaling #45893

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
bf0decd
fixed replication group scale down
sukumaar Jan 9, 2026
801eb24
test case naming convention
sukumaar Jan 9, 2026
5917685
changelog
sukumaar Jan 12, 2026
66be3f9
updated dowscaling logic and test cases
sukumaar Jan 16, 2026
966dd33
trivial change in comment
sukumaar Jan 23, 2026
6548739
Merge remote-tracking branch 'origin/main' into f-elasticache-rg-down…
YakDriver Jan 30, 2026
9a980a0
Fix lint issues
YakDriver Jan 30, 2026
559c377
fix: remove unnecessary trailing newline in replication_group.go
YakDriver Jan 30, 2026
1a16fb4
Make names clearer
YakDriver Jan 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions .changelog/45893.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,3 @@
```release-note:bug
resource/aws_elasticache_replication_group: Fixed an issue with downscaling `aws_elasticache_replication_group` when `cluster_mode="enabled"` and `num_node_groups` is reduced. Previously, downscaling could fail in certain scenarios; for example, if nodes `0001`, `0002`, `0003`, `0004`, and `0005` exist, and a user manually removes `0003` and `0005`, then sets `num_node_groups = 2`, terraform would attempt to delete `0003`, `0004`, and `0005`. This is now fixed, after this fix terraform will retrieve the current node groups before resizing.
```
39 changes: 29 additions & 10 deletions internal/service/elasticache/replication_group.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@ import (
"fmt"
"log"
"slices"
"sort"
"strconv"
"strings"
"time"
Expand Down Expand Up @@ -1258,19 +1259,37 @@ func modifyReplicationGroupShardConfigurationNumNodeGroups(ctx context.Context,
}

if oldNodeGroupCount > newNodeGroupCount {
// Node Group IDs are 1 indexed: 0001 through 0015
// Loop from highest old ID until we reach highest new ID
nodeGroupsToRemove := []string{}
for i := oldNodeGroupCount; i > newNodeGroupCount; i-- {
nodeGroupID := fmt.Sprintf("%04d", i)
nodeGroupsToRemove = append(nodeGroupsToRemove, nodeGroupID)
//scalling down scenario

Comment thread
sukumaar marked this conversation as resolved.
Outdated
nodeGroupIDs := []string{}
rg, err := findReplicationGroupByID(ctx, conn, d.Id())

if err != nil {
return fmt.Errorf("modifying ElastiCache Replication Group (%s) shard configuration: %w", d.Id(), err)
}
input.NodeGroupsToRemove = nodeGroupsToRemove
}

_, err := conn.ModifyReplicationGroupShardConfiguration(ctx, input)
for _, ng := range rg.NodeGroups {
if ng.NodeGroupId != nil {
nodeGroupIDs = append(nodeGroupIDs, *ng.NodeGroupId)
}
}
lenghtOfNodeGroupIDs := len(nodeGroupIDs)
Comment thread
sukumaar marked this conversation as resolved.
Outdated

if err != nil {
if lenghtOfNodeGroupIDs > newNodeGroupCount {
sort.Strings(nodeGroupIDs)

nodeGroupsToRemove := []string{}
reductionCount := oldNodeGroupCount - newNodeGroupCount

for i := lenghtOfNodeGroupIDs; i > (lenghtOfNodeGroupIDs - reductionCount); i-- {
nodeGroupsToRemove = append(nodeGroupsToRemove, nodeGroupIDs[i-1])
}

input.NodeGroupsToRemove = nodeGroupsToRemove
}
Comment thread
sukumaar marked this conversation as resolved.
Outdated
}

if _, err := conn.ModifyReplicationGroupShardConfiguration(ctx, input); err != nil {
return fmt.Errorf("modifying ElastiCache Replication Group (%s) shard configuration: %w", d.Id(), err)
}

Expand Down
255 changes: 255 additions & 0 deletions internal/service/elasticache/replication_group_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3370,6 +3370,207 @@ func TestAccElastiCacheReplicationGroup_Engine_Redis_LogDeliveryConfigurations_C
})
}

func TestAccElastiCacheReplicationGroup_RemoveNodeGroups_Valkey7(t *testing.T) {
ctx := acctest.Context(t)
if testing.Short() {
t.Skip("skipping long-running test in short mode")
}

var rg awstypes.ReplicationGroup
rName := acctest.RandomWithPrefix(t, acctest.ResourcePrefix)
resourceName := "aws_elasticache_replication_group.test"

acctest.ParallelTest(ctx, t, resource.TestCase{
PreCheck: func() { acctest.PreCheck(ctx, t) },
ErrorCheck: acctest.ErrorCheck(t, names.ElastiCacheServiceID),
ProtoV5ProviderFactories: acctest.ProtoV5ProviderFactories,
CheckDestroy: testAccCheckReplicationGroupDestroy(ctx, t),
Steps: []resource.TestStep{
{
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "valkey", "7.2", 5),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "valkey"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "7.2"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "5"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "5"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
{
ResourceName: resourceName,
ImportState: true,
ImportStateVerify: true,
ImportStateVerifyIgnore: []string{names.AttrApplyImmediately, "auth_token_update_strategy"},
},
{
PreConfig: func() {
conn := acctest.ProviderMeta(ctx, t).ElastiCacheClient(ctx)
timeout := 40 * time.Minute
nodeGroupsToRemove := []string{"0003", "0005"}

if err := resourceReplicationGroupShardModifyRemoveNodes(ctx, conn, rName, 3, nodeGroupsToRemove, timeout); err != nil {
t.Fatalf("error removing nodes: %s", err)
}
},
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "valkey", "8.0", 3),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "valkey"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "8.0"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "3"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "3"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
},
})
}

func TestAccElastiCacheReplicationGroup_RemoveNodeGroups_Valkey8(t *testing.T) {
ctx := acctest.Context(t)
if testing.Short() {
t.Skip("skipping long-running test in short mode")
}

var rg awstypes.ReplicationGroup
rName := acctest.RandomWithPrefix(t, acctest.ResourcePrefix)
resourceName := "aws_elasticache_replication_group.test"

acctest.ParallelTest(ctx, t, resource.TestCase{
PreCheck: func() { acctest.PreCheck(ctx, t) },
ErrorCheck: acctest.ErrorCheck(t, names.ElastiCacheServiceID),
ProtoV5ProviderFactories: acctest.ProtoV5ProviderFactories,
CheckDestroy: testAccCheckReplicationGroupDestroy(ctx, t),
Steps: []resource.TestStep{
{
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "valkey", "8.0", 5),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "valkey"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "8.0"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "5"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "5"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
{
ResourceName: resourceName,
ImportState: true,
ImportStateVerify: true,
ImportStateVerifyIgnore: []string{names.AttrApplyImmediately, "auth_token_update_strategy"},
},
{
PreConfig: func() {
conn := acctest.ProviderMeta(ctx, t).ElastiCacheClient(ctx)
timeout := 40 * time.Minute
nodeGroupsToRemove := []string{"0003", "0005"}

if err := resourceReplicationGroupShardModifyRemoveNodes(ctx, conn, rName, 3, nodeGroupsToRemove, timeout); err != nil {
t.Fatalf("error removing nodes: %s", err)
}
},
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "valkey", "8.0", 3),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "valkey"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "8.0"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "3"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "3"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
},
})
}

func TestAccElastiCacheReplicationGroup_RemoveNodeGroups_Redis7(t *testing.T) {
ctx := acctest.Context(t)
if testing.Short() {
t.Skip("skipping long-running test in short mode")
}

var rg awstypes.ReplicationGroup
rName := acctest.RandomWithPrefix(t, acctest.ResourcePrefix)
resourceName := "aws_elasticache_replication_group.test"

acctest.ParallelTest(ctx, t, resource.TestCase{
PreCheck: func() { acctest.PreCheck(ctx, t) },
ErrorCheck: acctest.ErrorCheck(t, names.ElastiCacheServiceID),
ProtoV5ProviderFactories: acctest.ProtoV5ProviderFactories,
CheckDestroy: testAccCheckReplicationGroupDestroy(ctx, t),
Steps: []resource.TestStep{
{
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "redis", "7.1", 5),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "redis"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "7.1"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "5"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "5"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
{
ResourceName: resourceName,
ImportState: true,
ImportStateVerify: true,
ImportStateVerifyIgnore: []string{names.AttrApplyImmediately, "auth_token_update_strategy"},
},
{
PreConfig: func() {
conn := acctest.ProviderMeta(ctx, t).ElastiCacheClient(ctx)
timeout := 40 * time.Minute
nodeGroupsToRemove := []string{"0003", "0005"}

if err := resourceReplicationGroupShardModifyRemoveNodes(ctx, conn, rName, 3, nodeGroupsToRemove, timeout); err != nil {
t.Fatalf("error removing nodes: %s", err)
}
},
Config: testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName, "valkey", "8.0", 3),
Check: resource.ComposeAggregateTestCheckFunc(
testAccCheckReplicationGroupExists(ctx, t, resourceName, &rg),
resource.TestCheckResourceAttr(resourceName, names.AttrEngine, "valkey"),
resource.TestCheckResourceAttr(resourceName, names.AttrEngineVersion, "8.0"),
resource.TestCheckResourceAttr(resourceName, "multi_az_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "automatic_failover_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "node_group_configuration.#", "3"),
resource.TestCheckResourceAttr(resourceName, "num_node_groups", "3"),
resource.TestCheckResourceAttr(resourceName, "replicas_per_node_group", "0"),
resource.TestCheckResourceAttr(resourceName, "cluster_enabled", acctest.CtTrue),
resource.TestCheckResourceAttr(resourceName, "at_rest_encryption_enabled", acctest.CtFalse),
resource.TestCheckResourceAttr(resourceName, "transit_encryption_enabled", acctest.CtFalse),
),
},
},
})
}

func testAccCheckReplicationGroupExists(ctx context.Context, t *testing.T, n string, v *awstypes.ReplicationGroup) resource.TestCheckFunc {
return func(s *terraform.State) error {
rs, ok := s.RootModule().Resources[n]
Expand Down Expand Up @@ -3673,6 +3874,15 @@ func testAccCheckResourceTags(resourceName string, kvs []kvp) []resource.TestChe
return checks
}

func resourceReplicationGroupShardModifyRemoveNodes(ctx context.Context, conn *elasticache.Client, replicationGroupID string, newNodeGroupCount int, nodeGroupsToRemove []string, timeout time.Duration) error {
return resourceReplicationGroupShardModify(ctx, conn, timeout, &elasticache.ModifyReplicationGroupShardConfigurationInput{
ApplyImmediately: aws.Bool(true),
NodeGroupCount: aws.Int32(int32(newNodeGroupCount)),
ReplicationGroupId: aws.String(replicationGroupID),
NodeGroupsToRemove: nodeGroupsToRemove,
})
}

func resourceReplicationGroupDisableAutomaticFailover(ctx context.Context, conn *elasticache.Client, replicationGroupID string, timeout time.Duration) error {
return resourceReplicationGroupModify(ctx, conn, timeout, &elasticache.ModifyReplicationGroupInput{
ReplicationGroupId: aws.String(replicationGroupID),
Expand Down Expand Up @@ -3723,6 +3933,22 @@ func resourceReplicationGroupModify(ctx context.Context, conn *elasticache.Clien
return nil
}

func resourceReplicationGroupShardModify(ctx context.Context, conn *elasticache.Client, timeout time.Duration, input *elasticache.ModifyReplicationGroupShardConfigurationInput) error {
_, err := conn.ModifyReplicationGroupShardConfiguration(ctx, input)
if err != nil {
return fmt.Errorf("error requesting modification: %w", err)
}

const (
delay = 30 * time.Second
)
_, err = tfelasticache.WaitReplicationGroupAvailable(ctx, conn, aws.ToString(input.ReplicationGroupId), timeout, delay)
if err != nil {
return fmt.Errorf("error waiting for modification: %w", err)
}
return nil
}

func formatReplicationGroupClusterID(replicationGroupID string, clusterID int) string {
return fmt.Sprintf("%s-%03d", replicationGroupID, clusterID)
}
Expand All @@ -3742,6 +3968,35 @@ resource "aws_elasticache_replication_group" "test" {
`, rName, engine)
}

func testAccReplicationGroupConfig_clusterModeWithNumNodeGroups(rName string, engine string, engineVersion string, numNodeGroups int) string {
return testAccReplicationGroupConfig_clusterModeWithNumNodeGroupsReplica(rName, engine, engineVersion, numNodeGroups, 0)
}
func testAccReplicationGroupConfig_clusterModeWithNumNodeGroupsReplica(rName string, engine string, engineVersion string, numNodeGroups int, replicationPerNodeGroup int) string {
return fmt.Sprintf(`
resource "aws_elasticache_replication_group" "test" {
replication_group_id = %[1]q
description = "test description"
node_type = "cache.t3.small"
port = 6379
apply_immediately = true
maintenance_window = "tue:06:30-tue:07:30"
snapshot_window = "01:00-02:00"
engine = %[2]q
engine_version = %[3]q
cluster_mode = "enabled"
num_node_groups = %[4]d
replicas_per_node_group = %[5]d

automatic_failover_enabled = true
multi_az_enabled = false

at_rest_encryption_enabled = false
transit_encryption_enabled = false

}
`, rName, engine, engineVersion, numNodeGroups, replicationPerNodeGroup)
}

func testAccReplicationGroupConfig_update_Valkey(rName string) string {
return fmt.Sprintf(`
resource "aws_elasticache_replication_group" "test" {
Expand Down
Loading