add storage_resize_mode mixed to opConfig CRD (#1947)

charts/postgres-operator/crds/operatorconfigurations.yaml

@@ -321,6 +321,7 @@ spec:
                     type: string
                     enum:
                       - "ebs"
+                      - "mixed"
                       - "pvc"
                       - "off"
                     default: "pvc"

charts/postgres-operator/values.yaml

@@ -196,7 +196,7 @@ configKubernetes:
   # whether the Spilo container should run with additional permissions other than parent.
   # required by cron which needs setuid
   spilo_allow_privilege_escalation: true
-  # storage resize strategy, available options are: ebs, pvc, off
+  # storage resize strategy, available options are: ebs, pvc, off or mixed
   storage_resize_mode: pvc
   # pod toleration assigned to instances of every Postgres cluster
   # toleration:

docs/reference/operator_parameters.md

@@ -486,10 +486,10 @@ configuration they are grouped under the `kubernetes` key.
 * **storage_resize_mode**
   defines how operator handles the difference between the requested volume size and
     the actual size. Available options are:
-    1. `ebs` : operator resizes EBS volumes directly and executes `resizefs` within a pod
+    1. `ebs`   : operator resizes EBS volumes directly and executes `resizefs` within a pod
-    2. `pvc` : operator only changes PVC definition
+    2. `pvc`   : operator only changes PVC definition
-    3. `off` : disables resize of the volumes.
+    3. `off`   : disables resize of the volumes.
-    4. `mixed` :operator  uses AWS API to adjust size, throughput, and IOPS, and calls pvc change for file system resize
+    4. `mixed` : operator uses AWS API to adjust size, throughput, and IOPS, and calls pvc change for file system resize
     Default is "pvc".
 
 ## Kubernetes resource requests

docs/user.md

@@ -1029,9 +1029,9 @@ specified but globally disabled in the configuration. The
 
 ## Increase volume size
 
-Postgres operator supports statefulset volume resize if you're using the
+Postgres operator supports statefulset volume resize without doing a rolling
-operator on top of AWS. For that you need to change the size field of the
+update. For that you need to change the size field of the volume description
-volume description in the cluster manifest and apply the change:
+in the cluster manifest and apply the change:
 
 ```yaml
 spec:
@@ -1040,22 +1040,29 @@ spec:
 ```
 
 The operator compares the new value of the size field with the previous one and
-acts on differences.
+acts on differences. The `storage_resize_mode` can be configured. By default,
+the operator will adjust the PVCs and leave it to K8s and the infrastructure to
+apply the change.
 
-You can only enlarge the volume with the process described above, shrinking is
+When using AWS with gp3 volumes you should set the mode to `mixed` because it
-not supported and will emit a warning. After this update all the new volumes in
+will also adjust the IOPS and throughput that can be defined in the manifest.
-the statefulset are allocated according to the new size. To enlarge persistent
+Check the [AWS docs](https://aws.amazon.com/ebs/general-purpose/) to learn
-volumes attached to the running pods, the operator performs the following
+about default and maximum values. Keep in mind that AWS rate-limits updating
-actions:
+volume specs to no more than once every 6 hours. 
 
-* call AWS API to change the volume size
+```yaml
+spec:
+  volume:
+    size: 5Gi # new volume size
+    iops: 4000
+    throughput: 500
+```
 
-* connect to pod using `kubectl exec` and resize filesystem with `resize2fs`.
+The operator can only enlarge volumes. Shrinking is not supported and will emit
-
+a warning. However, it can be done manually after updating the manifest. You
-Fist step has a limitation, AWS rate-limits this operation to no more than once
+have to delete the PVC, which will hang until you also delete the corresponding
-every 6 hours. Note, that if the statefulset is scaled down before resizing the
+pod. Proceed with the next pod when the cluster is healthy again and replicas
-new size is only applied to the volumes attached to the running pods. The
+are streaming.
-size of volumes that correspond to the previously running pods is not changed.
 
 ## Logical backups
 

manifests/operatorconfiguration.crd.yaml

@@ -319,6 +319,7 @@ spec:
                     type: string
                     enum:
                       - "ebs"
+                      - "mixed"
                       - "pvc"
                       - "off"
                     default: "pvc"

pkg/apis/acid.zalan.do/v1/crds.go

@@ -1428,6 +1428,9 @@ var OperatorConfigCRDResourceValidation = apiextv1.CustomResourceValidation{
 									{
 										Raw: []byte(`"ebs"`),
 									},
+									{
+										Raw: []byte(`"mixed"`),
+									},
 									{
 										Raw: []byte(`"pvc"`),
 									},