Chapter 2. Known and fixed issues
Learn about known issues for Data Grid Operator and find out which issues are fixed.
2.1. Known issues with Data Grid Operator deployments
This release does not include any known issues that affect Data Grid clusters that you manage with Data Grid Operator. For complete details about Data Grid, see the Data Grid 8.4 release notes.
2.2. Fixed in Data Grid Operator 8.4.14
Data Grid Operator 8.4.14 includes the following notable fixes:
2.3. Fixed in Data Grid Operator 8.4.12
Data Grid Operator 8.4.12 includes the following notable fixes:
-
JDG-6544 Changing the service type from
Cache
toDataGrid
does not result in automatic reconciliation -
JDG-6574 Setting up a custom
NodePort
value for cross-site replication has no effect - JDG-6573 Deadlock occurs when creating caches with a zero-capacity node and a single stateful node
-
JDG-6425 Updates to
podTargetLabel
are not reflected in pods
2.4. Fixed in Data Grid Operator 8.4.11
Data Grid Operator 8.4.11 includes the following notable fixes:
-
JDG-6584
Infinispan
CR should fail if Encryption keystore/certs missing in Secret
2.5. Fixed in Data Grid Operator 8.4.9
Data Grid Operator 8.4.9 includes the following notable fixes:
- JDG-6549 Data Grid Server HEAD requests failing with End Of File (EOF)
2.6. Fixed in Data Grid Operator 8.4.8
Data Grid Operator 8.4.8 includes the following notable fixes:
2.7. Fixed in Data Grid Operator 8.4.6
Data Grid Operator 8.4.6 includes the following notable fixes:
- JDG-6128 Data Grid Operator logs error message with stacktrace multiple times after cluster restart
- JDG-6127 Data Grid Operator repeatedly logs error messages indicating missing secrets while waiting for OpenShift to create a secret
-
JDG-6207
spec.Image
field is overwritten by Operand image for CVE releases -
JDG-6204
status.Operand.Image
field not updated when definingspec.Image
- JDG-6107 Creating caches with authorization fails and produces output issues
-
JDG-6055 Changing
memory
orcpu
values forspec.configListener
has no effect on the ConfigListener deployment -
JDG-5835 Scaling cluster down and up with
purge-on-startup=false
with one or more file-stores might result in stale entries
2.8. Fixed in Data Grid Operator 8.4.5
Data Grid Operator 8.4.5 includes the following notable fixes:
- JDG-6063 ConfigListener ignores the Cache CR metadata name when discovering existing Cache CRs
- JDG-5986 DNS discovery fails when pods are not marked as ready
- JDG-5935 Outdated names for metering labels
- JDG-5936 and JDG-5931 Metering labels are not updated after Data Grid Operator upgrade
- JDG-5939 Removal of the GracefulShutdownTask from Data Grid Operator
2.9. Fixed in Data Grid Operator 8.4.2
Data Grid Operator 8.4.2 includes the following notable fixes:
- JDG-5623 Gossip router fails to start with TLS configured on FIPS enabled OpenShift
-
JDG-5681 Incorrect
Cache
CR status when inmutable fields are modified -
JDG-5577
Cache
CR created bylistener
is not removed upon disabling the listener - JDG-5756 Users cannot configure LDAP because the authentication mechanisms of the default endpoint cannot be modified
-
JDG-5789
Infinispan.status.podStatus
field shows incorrect pod names for existing deployment topology -
JDG-5791 Data Grid Operator modifies the content of
spec.template
in theCache
CR after its creation -
JDG-5794
ConfigListener
log level cannot be modified -
JDG-5818 OpenShift rolling upgrades with
RollingMigration
strategy result inInfinispan
CR status asPending
- JDG-5820 Data Grid Operator fails when upgrading Data Grid Server from 8.3.1-1 to 8.4.0-x using the Hot Rod rolling migration strategy
-
JDG-5836
ConfigListener
stale CR check only compares resource names
2.10. Fixed in Data Grid Operator 8.4.0
Data Grid Operator 8.4.0 includes the following notable fixes:
- JDG-5680 Default Anti-affinity strategy configuration with the Data Grid Operator is not valid
-
JDG-5650
configListener
breaks with non-yamlCache
CRs template and large strings - JDG-5461 Server image doesn not enable Garbage collection (GC) logging by default
- JDG-5459 Zero controller execute can hang indefinitely if Zero Pod is not immediately ready