GridGain Enterprise Edition 8.5.6 Release Notes

Features and Improvements

  • Added version information in rolling update compatibility message. [#GG-13166]
  • Visor GUI: Disconnect from secure cluster if cluster was restarted or user idle. [#GG-14135]
  • Added login to "Visor connected" log message. [#GG-14745]
  • Scan query returns consistent results on topology changes. [#GG-14747]
  • Fixed potential deadlock on shutdown of reconnecting client node. [#GG-14771]
  • When client reconnects cluster, caches from its configuration are now recreated [#GG-14772]
  • Transaction label available in all nodes which participant in distributed transaction. [#GG-14776]
  • Web Console: Added missing DR properties on "Configuration" screen. [#WC-789]
  • Web Console: Updated landing page with new design. [#WC-907]

Fixed

  • Fixed assert error of on cancel command. [#GG-13944]
  • Attempt to PITR recovery with list of caches lead to exception. [#GG-14526]
  • Visor GUI: Fixed "Snapshots" tab refresh logic. [#GG-14633]
  • Output of all password arguments in snapshot-utility are masked. [#GG-14662]
  • Fixed bug: zero-sized partition file created in snapshot for an empty cache or cache group. [#GG-14708]
  • Fixed: Incremental snapshot fails if checkpointer writes pages of partition which was empty on snapshot start. [#GG-14733]
  • Added socket address to AuthenticationEvent. [#GG-14735]
  • Visor GUI: Fixed creation of snapshot in case when PITR is enabled. [#GG-14863]
  • Web Console: Fixed snapshot availability detection logic. [#WC-897]

GridGain Professional Edition 2.5.6 Release Notes

Features and Improvements

  • Failing nodes will be removed from topology as soon as possible on TcpCommunication initialization. Only server nodes can fail client in case of communication issues. [#IGNITE-7648]
  • Added coordinator information to control.sh output [#IGNITE-8894]
  • Web Console: Refactored all screens to show "no-data" message instead of blocking modal. [#IGNITE-10538]
  • Reduced checkpoint write lock hold time by optimistically saving metadata outside of checkpoint write lock [#IGNITE-10720]
  • Apply affinity changes on coordinator in multi-threaded mode [#IGNITE-10876]
  • Optimized heap utilization by history assignments [#IGNITE-10920]
  • Added discovery data to PluginProvider.validateNewNode, deprecate old method [#IGNITE-10944]
  • Web console: added more countries [#IGNITE-10961]
  • Improve join process, when marshaller's folder has many files. [#IGNITE-11023]
  • TcpDiscoverySpi discard queue heap usage optimization [#IGNITE-11058]
  • Data packet will compress on joining nodes to topology. [#IGNITE-11262]
  • Web Console: changed scaling on mobile devices [#IGNITE-11318]
  • Custom functionality of node fail through the jmx. [#IGNITE-11332]

Fixed

  • Node considered failed by cluster is forced to shut down when sending another Join Request. [#IGNITE-5569]
  • Added limit of diagnostic warnings [#IGNITE-6578]
  • Fixed possible failure during async partition cleanup [#IGNITE-9803]
  • Fixed indexes corruption on node stop under load [#IGNITE-10873]
  • Fix deserialization error of cache metrics with OptimizedMarshaller. [#IGNITE-10925]
  • Avoid to start WAL compressor in recovery mode. [#IGNITE-10928]
  • Fixed 'Invalid node order' assertion in discovery [#IGNITE-10935]
  • Web Agent: Implemented support for "--disable-demo". [#IGNITE-11096]
  • First node in new cluster makes additional filtering of cache configurations from old cluster [#IGNITE-11159]
  • Fixed PME hang when merged exchange node leaves before sending single message [#IGNITE-11204]
  • Create table with template doesn't work properly with data inserted from key-value API [#IGNITE-11219]
  • Healthy node ignores collection of failed nodes received from segmenting one [#IGNITE-11364]
  • Corrected client joining to topology. [#IGNITE-11454]
  • Fixed page store corruption that may be caused by early file downloader close during PITR [#IGNITE-11546]