OpsCenter 5.2.2 Release Notes

Release notes for the OpsCenter version 5.2.2 release.

Core 

  • Connection settings for monitored clusters and storage clusters are now distinct, which allows for different authentication/authorization/encryption configurations in the two clusters. See the [cassandra_storage] section in Cluster Configuration and the monitored_* properties in agent configuration if you are configuring address.yaml directly. (OPSC-6414)
  • Added a group_search_filter_with_dn parameter for LDAP configuration for LDAP environments that use a different search key in user_search_filter and group_search_filter. (OPSC-6187)
  • The default value for [ldap].opt_referrals has been changed from 0 to 1. This change should help some environments work more easily out of the box. (OPSC-6645)
  • The table-metrics.json file has been temporarily disabled in the diagnostics tarball while DataStax engineering troubleshoots some potential performance issues. (OPSC-6836)
  • Optimized auto-discovery of cluster topology that avoids making connections to multiple nodes on startup by taking advantage of CASSANDRA-9436. (OPSC-5943)
  • When using config file encryption, the config_encryption_key_path property in address.yaml no longer requires a trailing slash. (OPSC-5357)
  • Passwords can now contain all unicode characters when using DataStaxEnterpriseAuthentication. (OPSC-5920)
  • Tomcat log files from Search nodes are now included in the Diagnostics tarball. (OPSC-1858)
  • Fixed an issue where the agent would report as not connected after decommissioning a node and stopping the agent. If the agent is still running, there will still be a log message in opscenterd.log, but no warning in the UI. (OPSC-4723)
  • Fixed an issue where restarting a single datacenter after configuration would incorrectly display "All Nodes" in the Activity section. This was a display issue only. (OPSC-5238)
  • Fixed a typo in an error message when updating a cluster's Connection Settings to properly mention "saving cluster" instead of "deleting cluster". (OPSC-5243)
  • Fixed an issue when clicking Cancel on the Delete Cluster confirmation dialog. (OPSC-5701)
  • Fixed an issue that forced users to reopen the Move Node dialog after an invalid token was entered. (OPSC-5776)
  • Fixed an issue with the Add/Edit Cluster Connection Settings dialog where the Native Transport Port field was actually setting the deprecated Thrift Port field. (OPSC-6125)
  • Fixed an issue where a single agent would report as not connected when listen_address and rpc_address were configured differently. (OPSC-6659)

Monitoring 

  • Added new metrics for G1 Heap monitoring. (OPSC-5533)
  • Improved JMX connection handling on the agent to prevent false positives (reports of being UP) for down nodes, and to stop spamming agent.log when a node is down. (OPSC-4879)
  • Levels (e.g. , ALERT, ERROR) specified in alert notification configurations are now case-insensitive. (OPSC-6566)
  • Added a Clear Filters button to the List View when active filters display zero nodes in the results. (OPSC-5197)
  • Fixed an issue with the list of alerts extending beyond the bottom of the Alerts dialog. (OPSC-3334)
  • Fixed an issue causing a NullPointerException when collecting disk metrics on some Debian-based environments. (OPSC-3900)
  • Fixed issues with parsing output from the free command for collecting memory metrics in newer operating systems using procps-ng 3.3.10+. (OPSC-5045)
  • Fixed an issue where datacenter aggregate metrics would not display when a single node was bootstrapping. Bootstrapping nodes are now excluded from datacenter aggregates. (OPSC-5451)
  • Fixed an issue with percentile graphs not displaying "NO DATA" when empty. (OPSC-5884)
  • Fixed an issue within Dashboard widgets that prevented scrolling using the scrollbar within the Firefox web browser. (OPSC-5464)
  • Fixed linking directly to a Dashboard preset with an ampersand (&) in it when live_routes are enabled. (OPSC-6398)
  • Fixed an issue with incorrect rendering of the Ring View when resizing the ring after nodes go down. (OPSC-2366)
  • Fixed an issue with the node status not dynamically updating when hovering on a node in the Ring View. (OPSC-2446)
  • Fixed an issue with sorting tokens in the List View. (OPSC-3072)
  • Fixed a sorting issue with various statuses in List View. (OPSC-3073)
  • Fixed an issue with properly rounding the node Load value displayed to two decimal places within the Node Details dialog. (OPSC-5496)

Backup Service 

  • Improved Backup/Restore permissions behavior by replacing all uses of the bash "find" command when examining data directories and sstables. (OPSC-2682)
  • Added support for AWS S3 SSE (server-side encryption). This is configurable in the same dialog used to configure other S3 details. (OPSC-5932)
  • Added human-readable formatting to timestamps in some Point-in-Time Restore error messages. (OPSC-4409)
  • Added a warning when trying to start a restore when another restore process is already running. (OPSC-4547)
  • Made OOM related-errors during restore readable in the UI. (OPSC-5916)
  • Fixed an issue that prevented OpsCenter from restoring data when internode_encryption is enabled in cassandra.yaml. The path to cassandra.yaml is now passed to sstableloader using the -f argument. (OPSC-5670)
  • Fixed an issue with a missing scrollbar in the Configure Backup dialog. (OPSC-4184)
  • Fixed validation for the stream-throttling integer to disallow decimals. (OPSC-4390)
  • Fixed an issue that removed the On Server location when re-enabling commitlog archiving within a single UI session. (OPSC-5060)
  • Fixed an issue with automatically creating S3 buckets in the default region when running a backup. If buckets do not exist, they are now created automatically during location validation. (OPSC-6152)
  • Fixed an issue with schema.json being stored in an incorrect directory. This issue did not cause any user-facing issues with backup or restore. (OPSC-6217)
  • Fixed an issue with commitlog archiving where failures to properly archive files incorrectly returned as successful. (OPSC-6289)
  • Fixed an error that appeared when configuring commitlog archiving when it was already enabled. (OPSC-6305)
  • Fixed an issue with properly reporting the status of a PIT restore operation. (OPSC-6629)

Performance Service 

  • Added a confirmation step when enabling Performance Service objects in the Settings tab to ensure users are aware that doing so modifies dse.yaml. (OPSC-6044)
  • Fixed an issue with an unnecessary scrollbar displayed by Performance Service widgets in Firefox. (OPSC-5354)

Best Practice Service 

  • Replaced the "At least 2 seeds per DC" Best Practice rule with a more robust "Seed node configuration" rule. The new rule checks that all nodes use the same seed nodes, seed nodes use IPs instead of hostnames, and properly verifies that there are least two seeds specified per datacenter. (OPSC-6147)
  • The single Best Practice Service rule for Repair Service has been split into two distinct rules: a rule that checks for the Repair Service running, and another rule that compares the Repair Service time-to-complete with the lowest gc_grace setting across all tables. (OPSC-5135)

Repair Service 

  • Several typos fixed in log messages. (OPSC-6516)

Compatibility 

To see which versions of OpsCenter are compatible with the various Cassandra and DataStax Enterprise versions, see the OpsCenter Compatibility chart.