OpsCenter 6.0.4 Release Notes 

Release notes for the OpsCenter version 6.0.4 release.

Compatibility 

To see which versions of DataStax Enterprise are supported with OpsCenter 6.0, see the OpsCenter Compatibility chart.

For upgrade instructions, see the DataStax OpsCenter Upgrade Guide.

Known Issues

Important: Please be sure to review the list of Known Issues in OpsCenter 6.0 before running on a production cluster.

Core 

  • OpsCenter LDAP now filters multiple user LDAP groups by known OPSC roles if necessary during authentication. (OPSC-6946)
  • Improved a misleading error message for OpsCenter schema creation failures. (OPSC-9208)
  • Upgraded OpsCenter to use the latest version of the Cassandra Java Driver (3.1.0). (OPSC-9232)
  • When HTTPS is enabled, OpsCenter now issues cookies with the Secure flag. (OPSC-8713)
  • The Add Cluster dialog now requires cassandra keystore and agent keystore fields when client-to-node encryption is enabled. (OPSC-10150)
  • Appropriately cleaned up an issue with the tiered storage metric collector that caused OOM problems. (OPSC-10491)
  • Fixed a bug where a user's role was incorrectly updated during a password change by an admin user. (OPSC-8860)
  • Fixed an issue where OpsCenter could timeout while loading the events page. (OPSC-9762)
  • Fixed an issue where if client-to-node encryption was enabled for an OpsCenter monitored cluster and there was also a separate storage cluster, the agents were unable to connect to the storage cluster. (OPSC-10114)
  • Fixed an issue where the DataStax Agent would incorrectly bind its http server to broadcast_rpc_address for Cassandra when rpc_address was set to 0.0.0.0. Now the DataStax agent binds to whatever value rpc_address is set to unless it is overridden using the config parameter agent_rpc_interface in address.yaml. (OPSC-10243)
  • Fixed a bug where an improperly specified user_search_filter for LDAP would cause a stack trace to be logged that hid the proper LDAP error. (OPSC-10303)
  • Fixed a race condition when both an agent and its DSE node were down. The issue caused incorrectly reporting the node as up when the agent status changed. (OPSC-10336)
  • Fixed an issue that now allows sensitive config values to be specified in and properly decrypted using the agent address.yaml file. (OPSC-10396)
  • Fixed a bug in OpsCenter and the Datastax Agent where both systems could not authenticate with internal auth if it was specified as the secondary authentication method in DataStax Unified Authentication. (OPSC-10456)
  • Fixed an issue where the Spark Web UI proxy would be unusable when authentication was enabled. (OPSC-6606)
  • Fixed an issue that prevented proper startup of the HTTP reverse proxy to the Spark master. This issue only affected Analytics clusters running Spark. (OPSC-10221)

Monitoring 

  • Added the tombstones per read metric. (OPSC-2069)
  • The not_seen_threshold config value now also resets node health to unknown when it times out, in addition to marking agent connections as down. (OPSC-9844)
  • Improve memory use when many files are being streamed to or from Cassandra. (OPSC-10565)
  • Updated the ALLOW FILTERING description for the Best Practice Service rule as appropriate for certain analytic workloads. (OPSC-10297)
  • OpsCenter now supports named parameters for the Best Practices Service prepared statements rule check. (OPSC-7323)
  • Lowered the default for the slow_query_fetch_size configuration option available in address.yaml to alleviate potential stack overflows. (OPSC-7471)
  • Fixed a bug in the old metrics api when returning device-specific metrics. (OPSC-8885)
  • Fix 400 error with cluster-metrics API calls. (OPSC-9392)
  • Fixed a bug for Read and Write Latency graphs that caused no data to be displayed when multiple nodes were selected. (OPSC-9568)
  • Fixed a bug where the default table metric was not selected when View Metrics was clicked in the Data workspace. Added validation to the Add Metric dialog that warns the user a table selection is ignored if a non-table metric is selected. (OPSC-9766)
  • Fixed a bug where users were repeatedly informed about unavailable metrics in dashboard presets. (OPSC-10040)
  • Fixed an issue when restarting agents that caused the agent to lose track of the list of configured alerts. (OPSC-10579)
  • Fixed a rendering issue with the vnode ring view. (OPSC-9212)
  • Fixed an issue where schedules for the Best Practice service were re-enabled after an OpsCenter restart if all of the rules were disabled in the previous session. (OPSC-10157)
  • Fixed Solr filterCache best practice rule to account for DSE SolrFilterCache class. The best practice for the SolrFilterCache is to have a highWaterMarkMB setting of 256 (MB). (OPSC-6994)

Backup Service 

  • Improved the error message and its presentation for certain backup failures cases. (OPSC-9989)
  • Improved feedback when OpsCenter does not have write permissions to an S3 bucket. OpsCenter now presents an error if the agents fail to upload a backup to S3 due to permission errors. (OPSC-7029)
  • Clicking save on the scheduled backup job dialog now waits until the job has been successfully saved before closing the dialog. (OPSC-9607)
  • Cleaned up dirwatch threads when reconnecting to the Opscenter daemon. (OPSC-10528)
  • Fixed a bug that caused excessive logging about unexpected files encountered during the backup staging process. (OPSC-4415)
  • Fixed an issue with the Backup Report and status if an S3 destination of a scheduled backup is deleted before the backup has fully completed. (OPSC-6712)
  • Fixed an issue preventing point in time restore if the commit logs were archived to multiple destinations and at least one destination enabled compression. (OPSC-10293)
  • Fixed an issue causing errors when backing up DSE 5.0+ nodes that were still using an older (pre-DSE 5.0) data directory structure/format. (OPSC-10309)
  • Fixed a regression where backups did not appear in the restore dialog. (OPSC-10668)

Lifecycle Manager (LCM) 

  • Invalid package repository credentials in LCM now have a user-friendly error message. (OPSC-6554)
  • The meld remote executor used during LCM jobs has changed format from a .py text file to a .pyz zip application. This has no customer-facing impact except in cases where sudoers has required customization to allow meld to execute. (OPSC-7612)
  • Improved the error message when the DSE package cannot be found in the package repository during an LCM job. (OPSC-8391)
  • Editing an LCM config profile and navigating away without saving now warns users of unsaved changes. (OPSC-8424)
  • Password fields under LCM config profiles are now password input types. (OPSC-10379)
  • Reduced the likelihood of apt and yum corruption when terminating LCM jobs. (OPSC-9453)
  • LCM database operations resulting in a "database is locked" error will be retried several times before failing. (OPSC-9567)
  • LCM does not display clusters for importing if the clusters are not available for monitoring. (OPSC-9625)
  • Added SSH configuration options to alleviate connection issues when running an LCM job. (OPSC-9652)
  • Enabled LCM to downgrade agent packages, which prevents job errors when installing new versions of DSE using an older version of OpsCenter. (OPSC-9749)
  • Improved the ability of LCM to scale to larger job sizes. (OPSC-10493)
  • Fixed a positioning issue with action menus on the Clusters workspace of LCM. (OPSC-8988)
  • Fixed an issue that caused LCM clusters (in the Clusters workspace pane) to report that their most recently executed job was successful even though it had failed. (OPSC-9254)
  • Fixed an issue with the auto_bootstrap setting when importing clusters into LCM. The issue caused restarting Cassandra to fail after running a configure job. (OPSC-10151)
  • Fix rendering bug affecting some config profile values. (OPSC-10215)
  • Fixed a LCM job error when the meld remote executor takes too long to start up. (OPSC-10461)

Repair Service 

  • The Repair Service now continues to run when nodes go offline. (OPSC-6452)
  • If the Repair Service is unable to complete the repair operation within the time to completion window, the Services page now presents a link to the documentation on tuning the Repair Service in the informational status tooltip. (OPSC-6458)