Releases: FoundationDB/fdb-kubernetes-operator
Releases · FoundationDB/fdb-kubernetes-operator
v1.50.0
v1.50.0
Changes
Operator
- Fix the case where the process group gets removed without the addresses being included #2147
- Update the docs for e2e tests with a custom FDB version #2146
- Fix output for kubectl fdb exec command #2145
- Update docs unified image #2143
- Fix some bugs in the e2e test suite when only one version mapping is provided and fix the setup of the data loader #2141
- Add an operator e2e exclusion tests when a single log process has high latency #2138
v1.49.0
v1.48.0
v1.47.0
v1.47.0
Changes
Operator
- Add test suite name into namespace and cluster name to reduce conflicts #2126
- Refactor exec code to share same code and add plugin e2e test suite #2122
- Initial docs for coordinator recovery and e2e test #2121
- Read the coordinator IPs from the running Pods if they are present #2119
- Only release the lock when the cluster is reconciled #2117
- Improve logging for fdbclient and lock client to see what subreconciler issued the command #2116
v1.46.0
v1.45.0
v1.45.0
Changes
Operator
- Bug fix: Added required field namespace for global mode #2112
- Allow the operator to replace process groups with an I/O error #2111
- Make sure the operator provides an error message when the command line is too long #2110
- Ensure that the status checker unsets the FDB network options #2109
- Correct the base image for the unified image and increase the memory for the FDB pods to 8GiB #2107
v1.44.0
v1.44.0
Changes
Operator
v1.43.0
v1.43.0
Changes
Operator
- Move well-known env vars and config map keys into constants #2101
- Improve team tracker checks when DD was restarted #2100
- Try to get the running version based on the reachable coordinators during an upgrade #2098
- Check if the namespace is created by the same test suite #2097
- Make sure to only check the isolate process group annotation if a pod was returned #2095
- Update go version to 1.22.5 #2094
- Adding permissions for "nodes" resources to the helm chart (#2091) #2093
- Fix race condition in e2e test suite when checking if a pod is deleted #2092
- Run the change coordinator command before excluding the coordinator #2083
v1.42.0
v1.42.0
Changes
Operator
- Add more logging for missing processes in upgrade variations #2087
- Disable unified image tests until image is updated #2085
- Fix a bug where a process group with the same ID could be generated #2084
- Minor fixes for our e2e tests to improve reliability #2082
- Make sure the fdb-kubernetes-operator is forcing the ownership of the status subresource #2081
- Increase the replacement wait time #2079
- Bump github.com/hashicorp/go-retryablehttp from 0.7.2 to 0.7.7 #2078
- Remove enable node watch setting as this can be enabled with an env variable #2077
- Add an upgrade test case where a single fdb-kubernetes-monitor is not able to connect to the Kubernetes API #2076
- Refactor operator to make use of the fdb-kubernetes-monitor API #2075
- Minor bug fixes for the e2e test suite #2074
- Don't reset the last reconciled generation #2073
- Randomize process group ids #2072
- Make sure to wait until the namespace is completely deleted #2070
- Add better error message when command failed and remove flaky check #2069
- Add support for isolating a process group in the operator #2068
- Add future defaults for operator 2.0 release #2066
- Document the delay shutdown feature and check in update Pods if pod is missing #2065
- Add a test case for a multi-region cluster where the majority of coordinators fail #2064
- use extendEnv instead of append to container.Env #2062
- Make use of unstructured when using SSA to prevent default values causing issues #1962
v1.41.0
v1.41.0
Changes
Operator
- Change the image type from an env variable to an annotation #2063
- Set log dir and log group through environment variables for unified image #2060
- Add e2e test for HA cluster to make sure coordinators are changed #2058
- Fix e2e upgrade tests with chaos for unified image #2057
- Fix the calculation of expected processes during upgrades for multiple servers per pod #2056
- Make sure we update the correct process class #2055
- Make sure we skip the automatic replacements if the cluster is unavailable #2054
- Update the go version to 1.22.4 #2053
- Make sure to retry setting the process group under maintenance #2052
- Unified image code clean up and make use of the same constant #2051
- Disable availability check for upgrade test #2050
- Add e2e migration test from split to unified image #2048
- Add wait time for CI runs to reduce congestion #2046
- Prefer processes in coordinator selection for multi-region clusters #2045
- Change the useUnifiedImage setting to a more generic imageType #2042
- Add support for the node watch feature from the unified image #2040