OpsCenter 6.1.0 Release Notes 

Release notes for the OpsCenter version 6.1.0 release.

Highlights 

  • Support for DSE 5.1.
  • Revamped Repair Service! Smarter and more performant repair with real visibility into the progress and results of the repair process.
  • DataStax agent improvements: Backward compatibility with the installed OpsCenter version, which provides flexibility with the timing of upgrading agents; improved Agent Status view; and access to the DataStax Agent API in a Swagger UI console.

Be sure to check out New features for more details.

Upgrade Information 

Important: Please be sure to review the 6.1 changes pertinent to the release as noted in the DataStax OpsCenter Upgrade Guide. Configuration and other notable changes are provided in detail.

Compatibility 

To see which versions of DataStax Enterprise are supported with OpsCenter 6.x, see the OpsCenter Compatibility chart.

Known Issues 

Important: Please be sure to review the list of known issues before running a new OpsCenter version on a production DSE cluster.

Core 

  • The Hadoop workload type has been removed from the Datacenter dialog in LCM. Support for Hadoop has been deprecated. (OPSC-10466)
  • Changed criteria when the Analytics filter is used to constrain the nodes that provide data for a given graph. Previously only Hadoop nodes were considered when using this filter, but now only Spark nodes are considered. (OPSC-9562)
  • Added the swagger_enabled config option to address.yaml that exposes the DataStax Agent API in a Swagger UI console. (OPSC-10337)
  • Added an appropriate error when an incompatible route is called. (OPSC-9905)
  • Updated agent status to appropriately indicate incompatibility and limitations. (OPSC-9908)
  • If an agent's version will not permit a Best Practice Rule to run, a warning is shown when configuring the rule's schedule. (OPSC-9911)
  • OpsCenter added support for messaging latency metrics when monitoring DSE clusters from versions 5.1 and later. (OPSC-10093)
  • Added the ability to configure static JVM options in a jvm.options file within Lifecycle Manager config profiles for DSE clusters version 5.1.0 and later. The options were moved from cassandra-env.sh to jvm.options. See Configuring JVM options in LCM. (OPSC-10101)
  • The set of available cipher suites for TLS encryption of communication between the agent and DSE has been expanded to support the set of ciphers supported by the Java platform. This matches a similar change in Apache Cassandra™, which allows for negotiation of the strongest common cipher suite. (OPSC-10187)
  • Added support for DSE Integrated Authentication for JMX in DSE 5.1. See Configuring a JMX Connection using LCM. (OPSC-8752)
  • Added metrics for monitoring DSE Graph in OpsCenter. (OPSC-10180)
  • The banner that notifies users of agent issues no longer includes agent version mismatches that do not impact features. Incompatible version mismatches still appear in the banner. (OPSC-10327)
  • Agent Upgrade information is now presented in a text block above the agents status table. Agents can be upgraded by clicking the Upgrade Agents button. (OPSC-10615)
  • The Agent Version is now listed in the Agent Status tab. (OPSC-10616)
  • Agents with incompatible versions are now listed as problems in the agent status notification banner. (OPSC-10617)
  • Definitions files have been moved out of configuration directories (/etc/opscenter for package installs; install_location/conf for tarball installs.) The new locations are /var/lib/opscenter/definitions and install_location/definitions, respectively. This will be transparent to the majority of users. (OPSC-6704)
  • Added version prefix to agent API URIs. (OPSC-7531)
  • Removed path_to_installscript option from the [agents] section of opscenterd.conf. That option will now be ignored. The agent install process is now internal to Opscenter, and no external script is necessary. (OPSC-9119)
  • Moved agent-to-opscenter SSL settings from JVM_OPTS in the agent startup script to address.yaml. SSL settings in JVM_OPTS will continue to work as before if these settings are not specified in address.yaml. (OPSC-9398)
  • Added validation to os-metric/disk-space route. (OPSC-9928)
  • Enabled JVM heap dump by default for both opscenterd and the agent when there is an OutOfMemoryError. (OPSC-10861)
  • All memberOf attribute values are now evaluated during LDAP memberOf authentication. (OPSC-10893)
  • Included change details in audit log messages for updates to role permissions. (OPSC-10929)
  • Removed Clojure compiler warnings from agent startup. (OPSC-11068), (OPSC-11452)
  • Improved log messages when agents experience NFS read issues. (OPSC-11665)
  • Added validation of Content-Type header to agent API. (OPSC-10001)
  • No longer automatically install openjdk when directly installing OpsCenter from RPM or Deb packages. (OPSC-10260)
  • If OpsCenter authentication is enabled, users who do not have View Schema permission in their assigned role are not allowed to view schema data in the Data area of OpsCenter monitoring. OpsCenter presents an informational message that instructs users to contact their OpsCenter administrator for access. (OPSC-11319)
  • Fixed an issue which prevented some Solr operations from working correctly on multi-homed hosts. (OPSC-11437)
  • Fixed the /failover endpoint on the secondary OpsCenter instance. (OPSC-10720)
  • Fixed an issue which would cause opscenterd to mark an agent's http interface as down. (OPSC-10879)
  • Fixed a bug in agent lookup of datacenter name via JMX. This caused move token requests to fail during repairs. (OPSC-11788)
  • Fixed a potential SQL injection exploit in SQL authentication of OpsCenter (DatastaxEnterpriseAuth) in which an attacker could potentially gain access to the OpsCenter UI and its APIs using the password of any user that exists in the database. (OPSC-11991)

Monitoring 

  • The schema viewer in the Data area of the OpsCenter Monitoring UI now displays materialized views, secondary indexes, and user-defined types used on a table. Users with permissions can view query statements on the new User-Defined Types, User-Defined Functions, and User-Defined Aggregates tabs. (OPSC-4760)
  • Hover tooltips that describe each column in the Agent status view have been added. (OPSC-10769)
  • Added route /v1/monitor to agent to retrieve self-monitoring data on cassandra, monitored and storage clusters, jmx, rollups, messaging, component state, etc. (OPSC-3492)
  • Added convenient UI access to critical logs from within the Node details dialog. The logs are a labs feature that must be enabled. See Viewing logs from node details. (OPSC-9442)
  • Added LiveScannedHistogram metric to track the number of cells scanned during a read. (OPSC-10786)
  • Allow overriding content of http callback alerts. (OPSC-11054)
  • Alert badges in the global dashboard show an alerts list when clicked. (OPSC-4625)
  • Added support for the SSTables Repaired metrics at the node and table levels for DSE 5.1. (OPSC-10495)
  • Added Start Time column to the slow query log table in the Performance Service. (OPSC-10537)
  • Fixed a bug where best practice rule for Compaction Strategy would not log a failure. (OPSC-10987)

Lifecycle Manager (LCM) 

  • RLAC can now be enabled through LCM. (OPSC-10099)
  • LCM now allows jobs across clusters to run concurrently, but multiple jobs for the same cluster continue to run serially. (OPSC-8457)
  • LCM can no longer import clusters that have Hadoop enabled. Migrate to DSE Analytics prior to importing to LCM. (OPSC-10506)
  • LCM can no longer run jobs on clusters that have Hadoop enabled. (OPSC-10507)
  • The hive-site.xml file is now managed in the Spark section of the Configuration Profiles in LCM. (OPSC-10890)
  • Lifecycle Manager now validates cluster and datacenter names as ASCII strings instead of failing on non-ASCII characters during DSE installation. (OPSC-9276)
  • The DSE password now properly escapes the percentage (%) character. Users who escaped that character with the double %% workaround need to change their configuration back to a single %. (OPSC-9500)

Backup Service 

  • OpsCenter now warns users that attempt to restore backups without setting an explicit throttle value for stream throughput. Applies to DSE clusters from versions 4.8.7 and later. (OPSC-10185)
  • The [cloud] group and its only property, accepted_certs, are no longer part of OpsCenter config. (OPSC-11114)
  • OpsCenter switched to the Amazon SDK for S3 functionality. (OPSC-10033), (OPSC-10036)
  • Added support for backup and restore by datacenter. (OPSC-6155)
  • Re-implemented bandwidth throttling for S3 operations to make them more efficient and accurate. (OPSC-10073)
  • Added support for cloning from a backup using Local FS. (OPSC-6135)
  • Improved bucket name validation to more closely match what is allowed by AWS. (OPSC-6183)
  • Added the ability to alert on backup failure to the Event log for failed backups. (OPSC-7635)
  • The config parameter backup_file_queue_max has been removed from address.yaml. (OPSC-8045)
  • The Commit Log Archiving Consistency Best Practice Rule has been added to ensure all nodes within a cluster have Commit Log Archiving enabled to prevent any data loss during a PIT restore. (OPSC-10299)
  • Added support for a Local FS backup location and additional S3 or local locations in the Point in Time restore dialog in the Backup Service. (OPSC-10665), (OPSC-10728)
  • Tables with no data in them are now captured in backup data and can have their schemas restored. (OPSC-6558)
  • Performance improvements with local and S3 destinations. (OPSC-11727)
  • Updated the commit log archiving template to prevent an bug where commit logs would be transferred incompletely to S3. Users with commit log archiving currently active need to disable then enable commit log archiving for this change to take effect. (OPSC-11903)
  • Fixed the Unrecognized config key warning in the agent logs. (OPSC-9609)
  • Fixed an issue where materialized views were offered as tables that could be restored directly. (OPSC-11917)
  • Fixed a bug where a backup to a destination would be marked as successful even though the schema.json failed to upload and OpsCenter would be unable to restore from it. (OPSC-10815)

Repair Service 

  • Removed configuration settings from Repair Service that are no longer in use: ks_update_period, max_err_threshold, alert_on_repair_failure, repair_estimation_factor, incremental_range_repair. (OPSC-5895)
  • A new Status tab for the Repair Service conveys detailed information about subrange and incremental repairs. (OPSC-10020)
  • Removed the /<cluster ID>/services/repair/invalid_keyspaces API route since it is no longer used in the OpsCenter 6.1 Repair Service. (OPSC-11139)
  • The new Repair Service configuration option time_to_completion_target_percentage provides control over the speed with which repair jobs complete within the specified Time to completion parameter. See Adjusting or disabling the throttle for subrange repairs. (OPSC-2011)
  • Reworded log warning to use a less alarming phrase to describe an activation delay in the Repair Service. (OPSC-6453), (OPSC-8681)
  • Enhanced the Repair Service to ignore specific keyspaces and tables in addition to system keyspaces during a subrange repair. (OPSC-10458)
  • Excluded materialized views from incremental and subrange repairs. (OPSC-10955)
  • Clarified instructions for Time to completion and its relationship to gc_grace_seconds on the Repair Service settings page, plus provided links to relevant documentation. (OPSC-11454)
  • Removed the Repair Service progress bar from the Services page in lieu of a separate and much more detailed Repair Service status page. (OPSC-11560)
  • OpsCenter Services displays a Details link rather than a Configure link when services are already enabled. (OPSC-11661)
  • The target number of partitions per subrange can now be configured with the tokenranges_partitions configuration option. (OPSC-11801)
  • Added new Repair Service property tokenranges_http_timeout for setting a higher http timeout if necessary when retrieving tokenranges. (OPSC-11187)
  • Changed the system keyspaces excluded by the Repair Service to: dse_perf, system, system_distributed, system_schema, system_traces. (OPSC-11817)
  • Always now honor an API request to activate repair service. If a cluster is not yet in a stabilized state in which repairs can run, the Repair Service is activated in a paused state rather than raising an exception. (OPSC-11838)