Loading...

Type: Bug
Resolution: Unresolved
Priority: Major
Fix Version/s: None
Affects Version/s: None
Component/s: Operator
Labels:
None

Blocked:
False
Blocked Reason:
None
Ready:
False
CDW devel_ack:
CDW docs_ack:
CDW pm_ack:
CDW qa_ack:
CDW release:
Target Release:

RHDG BACKLOG GA
Intelligence Requested:
Market:

SFDC Cases Counter:
SFDC Cases Open:
SFDC Cases Links:

Issue only occurs when the ConfigListener is enabled.

Given an Infinispan CR that does not define XSite and the below Cache CR with xsite Backups configured, the expected behaviour is that upon Cache creation the Cache CR will have it's status set to Ready=False and a message containing error ISPN000571.

apiVersion: infinispan.org/v2alpha1
kind: Cache
metadata: 
  name: cache-xsite
spec: 
  clusterName: test-cache-cr
  name: cache-xsite
  template: |-
    <distributed-cache name="eu-customers" mode="ASYNC" statistics="true">
          <encoding media-type="application/x-protostream"/>
          <state-transfer enabled="false"/>
          <memory>
            <binary eviction="MEMORY" size="400000000"/> <!-- 400 MB -->
          </memory>
          <expiration lifespan="600000"/> <!-- 10 min -->
            <backups>
            <backup site="SiteA"
                    strategy="ASYNC">
              <state-transfer chunk-size="600"
                              timeout="2400000"
                              max-retries="30"
                              wait-time="2000"
                              mode="AUTO"/>
               <take-offline after-failures="5" min-wait="10000"/>
            </backup>
          </backups>
        </distributed-cache>
  updates: 
    strategy: retain

The current behaviour is that upon initial Cache creation on the server, the Cache CR correct reports the status:

apiVersion: infinispan.org/v2alpha1
kind: Cache
metadata: 
  annotations: 
    infinispan.org/listener-generation: "2"
  creationTimestamp: "2023-06-02T08:19:15Z"
  finalizers: 
  - finalizer.infinispan.org
  generation: 2
  labels: 
    test-name: TestCacheCR
  name: cache-xsite
  namespace: namespace-for-testing
  resourceVersion: "10226"
  uid: 36080a6e-432f-4f85-8f6b-a695148991a3
spec: 
  clusterName: test-cache-cr
  name: cache-xsite
  template: "distributedCache: \n  mode: \"ASYNC\"\n  statistics: \"true\"\n  backups: 
    \n    SiteA: \n      backup: \n        strategy: \"ASYNC\"\n        stateTransfer: 
    \n          chunkSize: \"600\"\n          timeout: \"2400000\"\n          maxRetries: 
    \"30\"\n          waitTime: \"2000\"\n          mode: \"AUTO\"\n        takeOffline: 
    \n          afterFailures: \"5\"\n          minWait: \"10000\"\n  encoding: \n
    \   mediaType: \"application/x-protostream\"\n  expiration: \n    lifespan: \"600000\"\n
    \ memory: \n    storage: \"BINARY\"\n    maxSize: \"400000000\"\n  stateTransfer: 
    \n    enabled: \"false\"\n"
  updates: 
    strategy: retain
status: 
  conditions: 
  - message: 'unable to create cache with template: unexpected HTTP status code (400):
      unexpected error creating cache, response: ISPN000571: RELAY2 not found in the
      protocol stack. Cannot perform cross-site operations.'
    status: "False"
    type: Ready

However, the ConfigListener then receives a CREATE cache event from the Server which causes the Cache CR to be reconfigured and the status to be reset to Ready=True.

apiVersion: infinispan.org/v2alpha1
kind: Cache
metadata: 
  annotations: 
    infinispan.org/listener-generation: "2"
  creationTimestamp: "2023-06-02T08:19:15Z"
  finalizers: 
  - finalizer.infinispan.org
  generation: 2
  labels: 
    test-name: TestCacheCR
  name: cache-xsite
  namespace: namespace-for-testing
  resourceVersion: "10227"
  uid: 36080a6e-432f-4f85-8f6b-a695148991a3
spec: 
  clusterName: test-cache-cr
  name: cache-xsite
  template: "distributedCache: \n  mode: \"ASYNC\"\n  statistics: \"true\"\n  backups: 
    \n    SiteA: \n      backup: \n        strategy: \"ASYNC\"\n        stateTransfer: 
    \n          chunkSize: \"600\"\n          timeout: \"2400000\"\n          maxRetries: 
    \"30\"\n          waitTime: \"2000\"\n          mode: \"AUTO\"\n        takeOffline: 
    \n          afterFailures: \"5\"\n          minWait: \"10000\"\n  encoding: \n
    \   mediaType: \"application/x-protostream\"\n  expiration: \n    lifespan: \"600000\"\n
    \ memory: \n    storage: \"BINARY\"\n    maxSize: \"400000000\"\n  stateTransfer: 
    \n    enabled: \"false\"\n"
  updates: 
    strategy: retain
status: 
  conditions: 
  - status: "True"
    type: Ready

The reason for this is that the cache get's created on the server, but only fails on startup, causing a Cache CREATED event to be triggered. The REST endpoint returns the error as expected and the Operator Cache CR controller correctly reports that the Cache failed and sets Ready=False. However, the ConfigListener then receives a Cache CREATED event via the ContainerResource#ConfigurationListener which sets Ready=True.

We should add an additional check to the Cache controller to inspect the health of a cache before setting Ready=True.

is related to

ISPN-14958 Add health endpoint per cache

Resolved

ISPN-14959 Allow Cache configurations to be evaluated against CacheManager state before Cache creation

Closed

Details

Description

Attachments

Issue Links

Easy Agile Planning Poker

Activity

People

Dates