Features and Improvements
- Ignite-Kafka Connector certified by Confluent: https://docs.gridgain.com/v8.4/docs/certified-kafka-connector [#GG-13027]
- Visor GUI: Improved usability on large cluster with many caches. [#GG-13195]
- PITR : handle situation when WAL got temporarily disabled for rebalancing [#GG-13652]
- Visor GUI: Added "Consistent ID" column to nodes panel. [#GG-13817]
- Fixed NPE in TTL manager on unwindEvicts. [#GG-13859]
- DR: dynamic caches support [#GG-13864]
- Node uptime metric now includes days count [#GG-13888]
- Fixed issue when calculation of operation estimation time was cut off earlier than it should (MOVE snapshot operation) [#GG-13890]
- Add a warning if license is expires in less than 90 days [#GG-13906]
- Visor GUI: Added functionality to preview and reset cache lost partitions. [#GG-13972]
- Web Console: Improved usability on large cluster with many caches. [#GG-14081]
- log4j 2x version was changed to 2.11.0 [#GG-14112]
- Jetty version updated to 9.4.11.v20180605 [#GG-14121]
- timestamp added to the name of log producing by IDLE_VERIFY to avoid previous logs overwriting [#GG-14204]
Fixed
- Snapshot utility: Added handling of missing folder in LIST command. [#GG-13758]
- Fixed issue during restore atomic caches [#GG-13866]
- Fix possible OOM on snapshot check operation [#GG-13899]
- Fixed NPE in DR on cluster activation [#GG-13901]
- Explicitly set queryEntities in StoredCacheData when applying changed configuration closure while restoring snapshot [#GG-13908]
- Fix error during delayed snapshot cancellation. [#GG-13912]
- Snapshot utility: Print to log correct user name in case of connecting to secured cluster. [#GG-14097]
- Snapshot utility: Do not output password to log and console. [#GG-14098]
- Visor GUI: Added option to show all caches if cluster configured with cache groups. [#GG-14145]
- REST API now requires ADMIN_OPS permission to activate/deactivate cluster [#GG-14147]
- Add support SSL connection in snapshot-utility.sh [#GG-14173]
- Snapshot utility: Print stack trace only to log file. [#GG-14232]
GridGain Professional Edition 2.5.3 Release Notes
Features and Improvements
- Visor CMD: Improved help for "start" command. [#IGNITE-1023]
- Web Console: Implemented responsive full screen layout. [#IGNITE-4010]
- Web Console: Configuration reworked to new modern design with improved usability. [#IGNITE-5466]
- Store TTL expire times in B+ tree on per-partition basis. [#IGNITE-5874]
- Added SslContextFactory.protocols and SslContextFactory.cipherSuites properties to configure encryption algorithms. [#IGNITE-6167]
- Web Console: Upgraded build to use Webpack 4. [#IGNITE-6816]
- Web Agent: Implemented support for comma-separated list of node URIs. [#IGNITE-7119]
- Fixed massive node drop from cluster on temporary network issues. [#IGNITE-7163]
- SQL: Added system view with topology nodes ("IGNITE.NODES") [#IGNITE-7700]
- SQL: Added system view with topology nodes attributes ("IGNITE.NODE_ATTRIBUTES") [#IGNITE-7701]
- Added ability to provide custom implementations of ExceptionListener for JmsStreamer. [#IGNITE-8154]
- Web Console: Implement execution of selected part of SQL query. [#IGNITE-8235]
- Web console: Added ability to show/hide password field value. [#IGNITE-8334]
- Web Console: Split "Sign In" page to three separate pages. [#IGNITE-8370]
- Increased dynamic cache starting speed [#IGNITE-8482]
- C++: Implemented thin C++ client [#IGNITE-8494]
- Fixed unexpected expired entry resurrection. [#IGNITE-8503]
- Visor GUI: Fixed display of rebalance state. [#IGNITE-8554]
- Turn system-critical Ignite threads into GridWorkers [#IGNITE-8562]
- Web Console: Added support for "collocated" query mode on Query screen. [#IGNITE-8568]
- validate_indexes command reports explicitly about entry which absent in data page [#IGNITE-8594]
- Add to control.sh tx utility information about transaction start time [#IGNITE-8601]
- Add support filter label not set in utility control.sh (control.sh --tx label null) [#IGNITE-8602]
- Added JMX-metrics: ClusterMetricsMXBean#getTotalBaselineNodes, ClusterMetricsMXBean#getActiveBaselineNodes. [#IGNITE-8603]
- Remove initOrder and loc keys from info in control.sh [#IGNITE-8620]
- Exposed thin client, JDBC and ODBC connections info through JMX [#IGNITE-8628]
- Failure processor dump state of all threads [#IGNITE-8642]
- Fixed performance issue if ExpiryPolicy is used with persistence. [#IGNITE-8681]
- Checkpoint logging was improved: useful debug logs promoted to info, walSegmentsCovered info calculation [#IGNITE-8737]
- Added transaction label to CU.txString() output [#IGNITE-8736]
- When topology changes and coordinator is also changed, we print log message like "Coordinator changed [prev=
gt;, cur= ;nodeid>gt;]". In a thread-safe way [#IGNITE-8738] ;nodeid> - Add methods to dump the ring structure and return current topology version [#IGNITE-8745]
- Exception in "no space left" situation is propagated to FailureHandler so it is able to take necessary action. [#IGNITE-8749]
- Optimized fsync system call over WAL segment. [#IGNITE-8761]
- Fixed nio-acceptor threads are indistinguishable in GridNioServer [#IGNITE-8781]
- The failureHandler handle error with meta-data [#IGNITE-8789]
- C++: Implemented Best Effort Affinity feature for the thin C++ client [#IGNITE-8794]
- Move transaction recovery logging to INFO level [#IGNITE-8798]
- Improved control.sh utility to show orphan local and remote transactions and ability to rollback them [#IGNITE-8808]
- Ease enabling experimental WAL commands in control.sh [#IGNITE-8815]
Fixed
- Fixed cache initialization event races. [#IGNITE-4939]
- Fixed failure detection in case of not receiving metrics updates from clients. [#IGNITE-5103]
- Improve cluster rebalancing routine if join#left node does not cause partition affinity assignment changes. [#IGNITE-7165]
- A service redeployment process during a topology change is fixed. [#IGNITE-7366]
- Web Console: Fixed demo for non-collocated joins. [#IGNITE-7927]
- Web Console: Fixed code generation for large numbers in configuration params. [#IGNITE-8140]
- Interrupting user thread during partition initialization will no longer cause node to stop. [#IGNITE-8203]
- The maximum value of the page counter of the B+Tree is increased. Also, failureHandler is added when the counter of the maximum number of retries overflows. [#IGNITE-8219]
- Fixed wrong RuntimeException handing on node stop. [#IGNITE-8238]
- Web Console: Fixed Dockerfile generation. [#IGNITE-8291]
- Fixed logging in LongJVMPauseDetector. [#IGNITE-8426]
- Web Console: Implemented connection to secured cluster. [#IGNITE-8428]
- Fix filter minSize in control.sh utility [#IGNITE-8467]
- Visor CMD: Added output of node "Consistent ID". [#IGNITE-8506]
- Fixed checkpoint history search for WAL delta rebalance [#IGNITE-8610]
- In case of client exchange being preempted from exchange history client nodes either try to reconnect or (with forceServerMode) shut down [#IGNITE-8657]
- Fixed PDS corruption when ExpiryPolicy is used. [#IGNITE-8659]
- WAL Iterator will stop and throws exception if an error occurs during read. [#IGNITE-8661]
- Display atomicity mode in control.sh utility output. [#IGNITE-8696]
- ODBC: Fixed ODBC crash in Linux when there is more than 1023 open file descriptors. [#IGNITE-8702]
- Corrected returning value of DataStorageMetrics.getTotalAllocatedSize [#IGNITE-8707]
- Fixed allocated pages metrics in case of partition move [#IGNITE-8706]
- Replace mislead U.warn implementation. [#IGNITE-8724]
- Fixed: EVT_CACHE_REBALANCE_PART_DATA_LOST event can be fired twice [#IGNITE-8746]
- Fixed entry expiration on entry updatedelete. [#IGNITE-8747]
- Fixed potential deadlock when trying to register new class metadata. [#IGNITE-8752]
- Fixed start of node outside of baseline with services configured. [#IGNITE-8754]
- Throw a correct error message when client optimized marshaller is trying to serialize too large object [#IGNITE-8755]
- In output of idle_verify added update counter for all conflicted partitions and information about partitions which rebalance [#IGNITE-8757]
- The default failure handler correct handle java.nio.file.AccessDeniedException [#IGNITE-8763]
- Fixed JVM crush during in-progress eviction and cache group stop [#IGNITE-8768]
- Daemon node isn't taken into account when switching cluster into compatibility mode [#IGNITE-8774]
- Fixed WAL rebalancing from backups [#IGNITE-8791]