Bug fixes:

• Fixed multi-threading model for proper handling or concurrent updates
• Fixed schema propagation logic to work with database names that require quoting

Other changes:

• Disabled query_thread_log by default (only applies to new instances)

Bug fixes:

• Fixed a bug with schema management queries failing at ClickHouse 20.3.+
• Fixed a bug with logging level configuration not being respected by metrics-exporter
• Fixed schema management logic to use ClickHouse pods directly bypassing LoadBalancer service
• Fixed a bug with host being deleted twice when deleting cluster or shards/replicas

Grafana dashboard has been also improved in this release.

Bug fixes:

• Sensitive data is masked now in operator logs
• Fixed a bug with replicated tables being deleted even when PVC has been retained (#314)
• Fixed a bug with annotations not being propagated to the pods (#289). Example 16-custom-labels-and-annotations-01.yaml

Documentation and examples has been updated to reflect recent changes.

Bug fixes:

• Fixed incorrect initialization of installation template in some cases

Improvements:

• Added podDistribution.scope in order to control scope of anti-affinity rules

Bug fixes:

• Fixed regression with Zookeeper configuration (#310)
• Fixed a bug in table_partitions metric export
• Removed excessive logging in metrics exporter

Improvements:

• Load balancer service now uses externalTrafficPolicy: Local by default
• ZooKeeper configuration is currently supported per cluster level (#209 via #288)

Bug fixes:

• Fixed a problem with upgrading operator for ClickHouseInstallations created by operator 0.8.0 and earlier
• Fixed a problem with updates to ClickHouse users configuration not always being propagated to ClickHouse servers (#282 )

Tweak:

• Added operator log level configuration

IMPORTANT NOTES:

1. Please upgrade operator via re-installation. There are change in resource definitions.
2. Operator versions 0.9.1 and 0.9.2 are not compatible with 0.9.0 and 0.9.3 in regards to upgrade. Clusters created with those versions have to be re-created.

Bug fixes:

• Fixed memory leak in metrics exporter (#176)

Bug fixes:

• Fixed regression with 'files' section of configuration (#261)
• Fixed discovery of monitored ClickHouse clusters by metrics exporter when it is being restarted (#239)

New features:

• Option to define template per ClickHouse replica (before it was only possible to have global, shard based or
• 'circular' replication shortcut (podDistribution.type: CircularReplication)
• hostTemplates in order to propagate custom ports across configuration
• hostNetwork support and portDistribution types. See examples starting from '14-' in examples folder.
  host based templates)
• New Grafana dashboard

New features:

• Introduced a set of 'podDistribution' affinity/anti-affinity rules to manage pod deployment in ClickHouse terms. See https://github.com/Altinity/clickhouse-operator/blob/master/docs/chi-examples/10-zones-03-advanced-03-pod-per-host-default-storage-class.yaml
• Default security configuration is improved:
  • All networks are disabled now. Localhost connections are enabled.
  • Connections from the cluster nodes are enabled via host_regexp.
  • ClickHouse user passwords are automatically hashed sha256 and stored as password_sha256_hex in config maps and files.
• All operator and user labels are propagated from ClickHouseInstallation to dependent objects: statefulsets, pods, services and configmaps
• Added some ClickHouse defaults that are automatically applied. It can be changed in operator configuration if necessary. See:
  • https://github.com/Altinity/clickhouse-operator/tree/master/config/config.d
  • https://github.com/Altinity/clickhouse-operator/tree/master/config/users.d
  • https://github.com/Altinity/clickhouse-operator/blob/master/config/config.yaml

Monitoring improvements:

• ClickHouse mutations are now available in monitoring
• Metrics fetch/error events were added
• Fixed a bug with monitoring events not properly labeled (also fixes #225)
• Fixed a bug when all hosts were dropped out of monitoring if one host fails

Other bug fixes:

• Fixed a bug when schema could not be created automatically on newer versions of ClickHouse (#227).
• Fixed a bug with zookeeper root entry not propagated to ClickHouse configuration

Upgrade notes:

• There were changes in ClickHouseInstallation CRD. It is recommended to remove and re-install the operator. Existing ClickHouse clusters will be picked up automatically.
• IMPORTANT: If you upgrade operator from 0.6.0 or earlier to 0.8.0 please make sure your ClickHouseInstallation name is shorter than 15 symbols. Otherwise DO NOT UPGRADE an operator.

Please contact [email protected] if you have upgrade issues or any questions.