chore: investigate rolling-update E2Es failures

.wordlist-en-custom.txt

@@ -123,6 +123,8 @@ EnterpriseDB's
 EnvFrom
 EnvFromSource
 EnvVar
+EphemeralVolumesSizeLimit
+EphemeralVolumesSizeLimitConfiguration
 ExternalCluster
 Fei
 Filesystem
@@ -366,6 +368,7 @@ TCP
 TLS
 TOC
 TODO
+TemporaryData
 TimelineId
 TopologyKey
 TopologySpreadConstraint
@@ -419,6 +422,7 @@ apiGroup
 apiGroups
 apiVersion
 apidoc
+apimachinery
 apis
 apiserver
 apparmor
@@ -623,6 +627,7 @@ endpointURL
 enterprisedb
 env
 envFrom
+ephemeralVolumesSizeLimit
 executables
 expirations
 extensibility
@@ -1077,6 +1082,7 @@ targetXID
 tbody
 tcp
 td
+temporaryData
 th
 thead
 timeLineID

api/v1/cluster_types.go

@@ -315,6 +315,10 @@ type ClusterSpec struct {
 	// +optional
 	Resources corev1.ResourceRequirements `json:"resources,omitempty"`
 
+	// EphemeralVolumesSizeLimit allows the user to set the limits for the ephemeral
+	// volumes
+	EphemeralVolumesSizeLimit *EphemeralVolumesSizeLimitConfiguration `json:"ephemeralVolumesSizeLimit,omitempty"`
+
 	// Name of the priority class which will be used in every generated Pod, if the PriorityClass
 	// specified does not exist, the pod will not be able to schedule.  Please refer to
 	// https://kubernetes.io/docs/concepts/scheduling-eviction/pod-priority-preemption/#priorityclass
@@ -429,6 +433,34 @@ const (
 	PhaseApplyingConfiguration = "Applying configuration"
 )
 
+// EphemeralVolumesSizeLimitConfiguration contains the configuration of the ephemeral
+// storage
+type EphemeralVolumesSizeLimitConfiguration struct {
+	// Shm is the size limit of the shared memory volume
+	Shm *resource.Quantity `json:"shm,omitempty"`
+
+	// TemporaryData is the size limit of the temporary data volume
+	TemporaryData *resource.Quantity `json:"temporaryData,omitempty"`
+}
+
+// GetShmLimit gets the `/dev/shm` memory size limit
+func (e *EphemeralVolumesSizeLimitConfiguration) GetShmLimit() *resource.Quantity {
+	if e == nil {
+		return nil
+	}
+
+	return e.Shm
+}
+
+// GetTemporaryDataLimit gets the temporary storage size limit
+func (e *EphemeralVolumesSizeLimitConfiguration) GetTemporaryDataLimit() *resource.Quantity {
+	if e == nil {
+		return nil
+	}
+
+	return e.TemporaryData
+}
+
 // ServiceAccountTemplate contains the template needed to generate the service accounts
 type ServiceAccountTemplate struct {
 	// Metadata are the metadata to be used for the generated

api/v1/cluster_types_test.go

@@ -18,6 +18,7 @@ package v1
 
 import (
 	corev1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/api/resource"
 	v1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/utils/ptr"
 
@@ -952,3 +953,21 @@ var _ = Describe("Cluster ShouldRecoveryCreateApplicationDatabase", func() {
 		Expect(result).To(BeFalse())
 	})
 })
+
+var _ = Describe("Ephemeral volume size limits", func() {
+	It("doesn't panic if the specification is nil", func() {
+		var spec *EphemeralVolumesSizeLimitConfiguration
+		Expect(spec.GetShmLimit()).To(BeNil())
+		Expect(spec.GetTemporaryDataLimit()).To(BeNil())
+	})
+
+	It("works correctly when fully specified", func() {
+		spec := &EphemeralVolumesSizeLimitConfiguration{
+			Shm:           ptr.To(resource.MustParse("10Mi")),
+			TemporaryData: ptr.To(resource.MustParse("20Mi")),
+		}
+
+		Expect(spec.GetShmLimit().String()).To(Equal("10Mi"))
+		Expect(spec.GetTemporaryDataLimit().String()).To(Equal("20Mi"))
+	})
+})

api/v1/cluster_webhook.go

@@ -987,6 +987,19 @@ func (r *Cluster) validateResources() field.ErrorList {
 		}
 	}
 
+	ephemeralStorageRequest := r.Spec.Resources.Requests.StorageEphemeral()
+	ephemeralStorageLimits := r.Spec.Resources.Limits.StorageEphemeral()
+	if !ephemeralStorageRequest.IsZero() && !ephemeralStorageLimits.IsZero() {
+		ephemeralStorageRequestGtThanLimit := ephemeralStorageRequest.Cmp(*ephemeralStorageLimits) > 0
+		if ephemeralStorageRequestGtThanLimit {
+			result = append(result, field.Invalid(
+				field.NewPath("spec", "resources", "requests", "storage"),
+				ephemeralStorageRequest.String(),
+				"Ephemeral storage request is greater than the limit",
+			))
+		}
+	}
+
 	return result
 }
 

api/v1/cluster_webhook_test.go

@@ -3236,6 +3236,38 @@ var _ = Describe("validateResources", func() {
 		Expect(errors[0].Detail).To(Equal("Memory request is greater than the limit"))
 	})
 
+	It("returns no error when the ephemeral storage request is correctly set", func() {
+		cluster.Spec.Resources.Requests["ephemeral-storage"] = resource.MustParse("1")
+		cluster.Spec.Resources.Limits["ephemeral-storage"] = resource.MustParse("1")
+
+		errors := cluster.validateResources()
+		Expect(errors).To(BeEmpty())
+	})
+
+	It("returns an error when the ephemeral storage request is greater than ephemeral storage limit", func() {
+		cluster.Spec.Resources.Requests["ephemeral-storage"] = resource.MustParse("2")
+		cluster.Spec.Resources.Limits["ephemeral-storage"] = resource.MustParse("1")
+
+		errors := cluster.validateResources()
+		Expect(errors).To(HaveLen(1))
+		Expect(errors[0].Detail).To(Equal("Ephemeral storage request is greater than the limit"))
+	})
+
+	It("returns three errors when CPU, Memory, and ephemeral storage requests are greater than limits", func() {
+		cluster.Spec.Resources.Requests["cpu"] = resource.MustParse("2")
+		cluster.Spec.Resources.Limits["cpu"] = resource.MustParse("1")
+		cluster.Spec.Resources.Requests["memory"] = resource.MustParse("2Gi")
+		cluster.Spec.Resources.Limits["memory"] = resource.MustParse("1Gi")
+		cluster.Spec.Resources.Requests["ephemeral-storage"] = resource.MustParse("2")
+		cluster.Spec.Resources.Limits["ephemeral-storage"] = resource.MustParse("1")
+
+		errors := cluster.validateResources()
+		Expect(errors).To(HaveLen(3))
+		Expect(errors[0].Detail).To(Equal("CPU request is greater than the limit"))
+		Expect(errors[1].Detail).To(Equal("Memory request is greater than the limit"))
+		Expect(errors[2].Detail).To(Equal("Ephemeral storage request is greater than the limit"))
+	})
+
 	It("returns two errors when both CPU and Memory requests are greater than their limits", func() {
 		cluster.Spec.Resources.Requests["cpu"] = resource.MustParse("2")
 		cluster.Spec.Resources.Limits["cpu"] = resource.MustParse("1")

config/crd/bases/postgresql.cnpg.io_clusters.yaml

@@ -1842,6 +1842,26 @@ spec:
                       x-kubernetes-map-type: atomic
                   type: object
                 type: array
+              ephemeralVolumesSizeLimit:
+                description: EphemeralVolumesSizeLimit allows the user to set the
+                  limits for the ephemeral volumes
+                properties:
+                  shm:
+                    anyOf:
+                    - type: integer
+                    - type: string
+                    description: Shm is the size limit of the shared memory volume
+                    pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
+                    x-kubernetes-int-or-string: true
+                  temporaryData:
+                    anyOf:
+                    - type: integer
+                    - type: string
+                    description: TemporaryData is the size limit of the temporary
+                      data volume
+                    pattern: ^(\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))(([KMGTPE]i)|[numkMGTPE]|([eE](\+|-)?(([0-9]+(\.[0-9]*)?)|(\.[0-9]+))))?$
+                    x-kubernetes-int-or-string: true
+                type: object
               externalClusters:
                 description: The list of external clusters which are used in the configuration
                 items:

docs/src/appendixes/object_stores.md

@@ -94,6 +94,17 @@ spec:
         [...]
 ```
 
+### S3 lifecycle policy
+
+Barman Cloud writes objects to S3, then does not update them until they are
+deleted by the Barman Cloud retention policy. A recommended approach for an S3
+lifecycle policy is to expire the current version of objects a few days longer
+than the Barman retention policy, enable object versioning, and expire
+non-current versions after a number of days. Such a policy protects against
+accidental deletion, and also allows for restricting permissions to the
+CloudNativePG workload so that it may delete objects from S3 without granting
+permissions to permanently delete objects.
+
 ### Other S3-compatible Object Storages providers
 
 In case you're using S3-compatible object storage, like **MinIO** or

docs/src/cloudnative-pg.v1.md

@@ -1546,6 +1546,14 @@ https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/
 for more information.</p>
 </td>
 </tr>
+<tr><td><code>ephemeralVolumesSizeLimit</code> <B>[Required]</B><br/>
+<a href="#postgresql-cnpg-io-v1-EphemeralVolumesSizeLimitConfiguration"><i>EphemeralVolumesSizeLimitConfiguration</i></a>
+</td>
+<td>
+   <p>EphemeralVolumesSizeLimit allows the user to set the limits for the ephemeral
+volumes</p>
+</td>
+</tr>
 <tr><td><code>priorityClassName</code><br/>
 <i>string</i>
 </td>
@@ -2160,6 +2168,38 @@ a Role in a PostgreSQL instance</p>
 
 
 
+## EphemeralVolumesSizeLimitConfiguration     {#postgresql-cnpg-io-v1-EphemeralVolumesSizeLimitConfiguration}
+
+
+**Appears in:**
+
+- [ClusterSpec](#postgresql-cnpg-io-v1-ClusterSpec)
+
+
+<p>EphemeralVolumesSizeLimitConfiguration contains the configuration of the ephemeral
+storage</p>
+
+
+<table class="table">
+<thead><tr><th width="30%">Field</th><th>Description</th></tr></thead>
+<tbody>
+<tr><td><code>shm</code> <B>[Required]</B><br/>
+<i>k8s.io/apimachinery/pkg/api/resource.Quantity</i>
+</td>
+<td>
+   <p>Shm is the size limit of the shared memory volume</p>
+</td>
+</tr>
+<tr><td><code>temporaryData</code> <B>[Required]</B><br/>
+<i>k8s.io/apimachinery/pkg/api/resource.Quantity</i>
+</td>
+<td>
+   <p>TemporaryData is the size limit of the temporary data volume</p>
+</td>
+</tr>
+</tbody>
+</table>
+
 ## ExternalCluster     {#postgresql-cnpg-io-v1-ExternalCluster}
 
 

docs/src/cluster_conf.md

@@ -35,6 +35,25 @@ spec:
 You can find a complete example using projected volume template to mount Secret and Configmap in
 the [cluster-example-projected-volume.yaml](samples/cluster-example-projected-volume.yaml) deployment manifest.
 
+## Ephemeral volumes
+
+CloudNativePG relies on [ephemeral volumes](https://kubernetes.io/docs/concepts/storage/ephemeral-volumes/)
+for part of the internal activities. Ephemeral volumes exist for the sole duration of
+a pod's life, without persisting across pod restarts.
+
+### Volume for temporary storage
+
+An ephemeral volume used for temporary storage. An upper bound on the size can be
+configured via the `spec.ephemeralVolumesSizeLimit.temporaryData` field in the cluster
+spec.
+
+### Volume for shared memory
+
+This volume is used as shared memory space for Postgres, also an ephemeral type but
+stored in-memory. An upper bound on the size can be configured via the
+`spec.ephemeralVolumesSizeLimit.shm` field in the cluster spec. This is used only
+in case of [PostgreSQL running with `posix` shared memory dynamic allocation](postgresql_conf.md#dynamic-shared-memory-settings).
+
 ## Environment variables
 
 Some system behavior can be customized using environment variables. One example is

docs/src/postgresql_conf.md

@@ -440,6 +440,16 @@ You should get something similar to the following output:
 shm on /dev/shm type tmpfs (rw,nosuid,nodev,noexec,relatime,size=******)
 ```
 
+If you would like to set a maximum size for the `shm` volume, you can do so by
+setting the `spec.ephemeralVolumesSizeLimit.shm` field in the `Cluster` resource.
+For example:
+
+```yaml
+spec:
+  ephemeralVolumesSizeLimit:
+    shm: 1Gi
+```
+
 ### System V shared memory
 
 In case your Kubernetes cluster has a high enough value for the `SHMMAX`

go.mod

@@ -29,7 +29,7 @@ require (
 	github.com/thoas/go-funk v0.9.3
 	go.uber.org/atomic v1.11.0
 	go.uber.org/zap v1.26.0
-	golang.org/x/exp v0.0.0-20231005195138-3e424a577f31
+	golang.org/x/exp v0.0.0-20231006140011-7918f672742d
 	golang.org/x/sys v0.13.0
 	gopkg.in/yaml.v3 v3.0.1
 	k8s.io/api v0.28.2
@@ -87,14 +87,14 @@ require (
 	github.com/xlab/treeprint v1.2.0 // indirect
 	go.starlark.net v0.0.0-20230525235612-a134d8f9ddca // indirect
 	go.uber.org/multierr v1.11.0 // indirect
-	golang.org/x/crypto v0.13.0 // indirect
-	golang.org/x/net v0.15.0 // indirect
+	golang.org/x/crypto v0.14.0 // indirect
+	golang.org/x/net v0.16.0 // indirect
 	golang.org/x/oauth2 v0.8.0 // indirect
-	golang.org/x/sync v0.3.0 // indirect
-	golang.org/x/term v0.12.0 // indirect
+	golang.org/x/sync v0.4.0 // indirect
+	golang.org/x/term v0.13.0 // indirect
 	golang.org/x/text v0.13.0 // indirect
 	golang.org/x/time v0.3.0 // indirect
-	golang.org/x/tools v0.13.0 // indirect
+	golang.org/x/tools v0.14.0 // indirect
 	gomodules.xyz/jsonpatch/v2 v2.4.0 // indirect
 	google.golang.org/appengine v1.6.7 // indirect
 	google.golang.org/protobuf v1.31.0 // indirect