Changelog
This section lists the changes that were made to GigaSpaces products for 15.8.
15.8.1
Features and Enhancements
- GS-14463: Kubernetes increase liveness probe values
- GS-14487: Enhance Java version parsing to support early access builds
- GS-14486: Add username to system report in beginning of log
- GS-14474: Support running Docker and Kubernetes as non-root user
- GS-14478: Improve Space clear with null template when Space contains broadcast table
- GS-14475: Support loading "com.gs.security.properties-file" from url
- GS-14476: Support loading "java.util.logging.config.file" from url
- GS-14477: Support loading "com.gigaspaces.metrics.config" from url
- GS-14429: Issue with deploying Manager to multiple namespaces in Kubernetes
- GS-14389: Move heap report to zookeeper
Resolved Issues
- GS-14479: Unable to change zookeeper clientPort
- GS-14465: Allow to configure security in Kubernetes
- GS-14409: Can't connect to a secured space using Ops Manager, same goes to the query data in the rest API
- GS-14460: Concurrent modification bug in system task
- GS-14421: Quiesce Mode during fail over is not resolved even after the curator client already reports RECONNECTED
- GS-14438: Default Cursor iterator does not support LRU space
- GS-14435: Default Cursor iterator does not support local view
- GS-14450: NPE running cursor iterator in DEBUG log level
- GS-14399: Upgrade from 14.5 to 15.x - ClusterInfo.getInstanceId() is wrong in some conditions
- GS-14442: "Templating dataTypeName" error in Grafana v7.4.1 when displaying Space Object Type Index Metrics
- GS-14427: Kubernetes operator - Pass LRMI properties to PU
- GS-14426: Kubernetes operator - Pass license to Kubernetes manager
- GS-14423: Marshalling exception caused by Oshi when parsing Mac OS version
- GS-14422: Block unauthorized features in REST and Ops-UI
- GS-14417: REST: Path variable with dot (.) is truncated in URL
- GS-14388: Upgrade from 12.3 to 15.5.1 log regression — unable to create log per PU using slf4j logger api and Log4j configuration
- GS-14406: Add scroll bar for data types and indexes table
- GS-14396: Query time shown in Ops Manager is incorrect
- GS-14401: Service card is green despite number of actual instances greater than planned
- GS-14393: Ops Manager doesn't display number of entries in single Space due to NPE
- GS-14394: Deploying/undeploying a Space succeeds but returns an operation timeout in the CLI
Known Issues
- GS-14542: Exception in manager: java.sql.SQLException: statement is invalid while deleting rows from tables
- GS-14544: Exception in manager: org.hsqldb.HsqlException: statement execution aborted: timeout reached
15.8
Features and Enhancements
- GS-14332: Add more configurations for deploying a processing unit with Kubernetes Operator
- GS-14369: Ops-UI : Expose broadcast object and storage optimized
- GS-14379: Kubernetes operator web app deployment
- GS-14342: Count number of read operations instead of number of actually read objects while handling data type read counts
- GS-14340: Support Horizontal Scale for Kubernetes Stateful Sets
- GS-14341: Support vertical scale for Kubernetes Stateful Sets using Operator
- GS-14350: Block scale in/out and undeploy features if the processing unit is deployed from helm charts
- GS-14329: Remove commons-lang and commons-codec from lib/platform/commons
- GS-14347: Upgrade Apache POI to 4.1.2
- GS-14348: Upgrade Snakeyaml to 1.27
- GS-14349: Upgrade Zookeeper’s Netty to 4.1.48
- GS-14273: ElasticGrid - Add service name to GSC log folder name in service dump
- GS-14274: Add more logs to service dump in ElasticGrid
- GS-14344: Upgrade Spring to 5.1.19
- GS-14345: Upgrade Spring Security to 5.1.13
- GS-14343: Upgrade Jetty to 9.4.35
- GS-14158: New gs.sh cli tool does not support deployment of 0 instances
- GS-14009: Deterministic deployment should use demote instead of restarts
- GS-14337: Ops-ui: reads by partition per Data type displays empty bar chart
- GS-14336: Remove unnecessary Space metrics metadata to decrease HSQLDB memory footprint
- GS-14335: RESTful Space query throws NullPointerException caused by non-required fields being empty
- GS-14326: Add support for deploy of Processing Unit using Kubernetes Operator
- GS-14328: Deploy stateless PU using Kubernetes Operator
- GS-14329: Remove dependency of commons-lang and commons-codec jars from lib/platform/commons
- GS-14310: JDBC - Support functions in Select clause
- GS-14311: Support Inner Join
- GS-14312: Support non-scalar subqueries (subquery as table)
- GS-14313: JDBC - Execute Join collocated with the data for relevant queries
- GS-14314: JDBC - Add support for Left Join
- GS-14315: Support function in Group By clause
- GS-14305: ElasticGrid - add PostgreSQL as a built-in service
- GS-14297: ElasticGrid — allow auto deploying blueprint when using data connectivity
- GS-14013: New configuration option to define initial port for JMX
- GS-14295: Upgrading Curator to release 5.1.0
- GS-14291: ElasticGrid - add option for disabling SSL on manager service
- GS-14285: Explain Plan is now available in Ops Manager
- GS-14242: ElasticGrid - improve life cycle of processing units
- GS-14277: ElasticGrid - allow public access for AWS Cluster
- GS-14267: ElasticGrid - use AWS credentials from the client host for node add/remove commands
- GS-14263: ElasticGrid - Use --server=<node-ip> to communicate with cluster
- GS-14264: ElasticGrid - ask for confirmation when running destroy command
- GS-14251: Define liveness health check and cancel restarts in readiness health check for stateful service on ElasticGrid
- GS-14258: Display a tooltip showing the index hits count in the object drill down page of the Ops Manager UI
- GS-14225: Added support for scale up/down stateful pu without HA
- GS-14247: Set logger level using environment variable or system property
- GS-14246: Set metrics xml config file via GS_METRICS_CONFIG environment variable
- GS-14245: Set metric sampler using environment variable or system property
- GS-14205: Upgrade Oshi to 5.2.5
- GS-14239: Change default pu memory to 512mb and managers default to 1024mb on ElasticGrid
- GS-14236: Fix text overflow in GSCTL output to display all details including the token
- GS-14219: Rename rest-manager to ops-manager in ElasticGrid list-services command
Resolved Issues
- GS-14334: ClassNotFoundException when running on Java 8 on an IBM J9 VM within WebSphere
- GS-14367: Ops-UI services page - Align IOPS chart to the other charts
- GS-14361: Scale up/down - show cpu units in millicores
- GS-14355: Ops-UI: Hide IOPS chart of non-stateful services
- GS-14357: SQL functions executed twice if function also in Group By
- GS-14331: CPU and memory data not displayed in ops-ui
- GS-14322: DB tables for the Ops Manager metrics are not created if local language is not English
- GS-14271: Backup should recover from primary Space when using non-persistent blob-store data policy
- GS-14307: ElasticGrid - Unable to access the data grid from outside the cluster
- GS-14308: Unable to login to Ops Manager when switching between secured and unsecured modes
- GS-14306: Ops Manager data connectivity assigns wrong index type
- GS-14048: Property com.gs.jmx.enabled to false does not disable JMX
-
GS-14282: Illegal state - connection to ZooKeeper resumed but leadership is held by another session
- GS-14284: ElasticGrid - undeploy of a failed deployment leaks resources and causes error on redeploy
- GS-14280: NullPointerException when failing to initialize connection to HSQLDB
- GS-14278: ElasticGrid - scale out may hang when adding more than one partition with insufficient resources
- GS-14279: ElasticGrid - node remove command may terminate nodes without draining them when removing multiple nodes
- GS-14252: ElasticGrid - create cluster may fail when user-provided VPC/SG/subnet is in use
- GS-14265: CPU utilization not consistently reported across platforms
- GS-14261: Stateful blueprints fail to compile using Java 11 due to missing javax.annotation dependency
- GS-14243: ElasticGrid - Creating AWS cluster using user-provided security group may fail when the inbound rules contain a rule for the client machine IP
- GS-14074: After upgrading to 15 version, classNotFound exception due to usage of symbolic links
- GS-14254: Order of processing removal events prevents proper removal of VirtualMachine from Machine object in Admin API
- GS-14238: Retry policy used to connect a client to ZooKeeper elapses prematurely
- GS-14223: Space remains in stopped state after network reconnection
- GS-14257: Ops Manager UI displays an empty index
- GS-14234: NPE while running Explain Plan for Select Query that has id in Where clause
- GS-14256: Change log level of SQLTransientConnectionException from SEVERE to WARNING
- GS-14206: Oshi cause permanent WMI requests from XAP processes that are leading to a high CPU consumption
- GS-14255: Ignore GS_PU_DYNAMIC_PARTITIONING environmental variable when deploying non-partitioned service
- GS-14240: Create command might fail when using auto-generated subnet because of wrong availability zone choice
- GS-14241: Grafana got 502 error in ElasticGrid caused by DNS resolution
- GS-14080: Can't change port and IP in WEB UI
- GS-14226: Change default disk size in ElasticGrid AWS
- GS-14233: On Space shutdown, close ZooKeeper client's connection state listener
- GS-14216: ElasticGrid support for Java 11 and later
- GS-14228: After gsm restarted, exception [java.sql.SQLException: statement is closed] thrown for gsc
- GS-14232: 'gsctl node add' command fails if a wrong component name was provided
- GS-14220: ElasticGrid node add does not work with profile name that includes spaces
- GS-14221: gsctl info command does not work if the cluster is not created yet
- GS-14217: ElasticGrid disabling metrics when metrics-related built-in services are disabled
Known Issues
For client version 14.0 or older, there is a possibility that the redo log will be corrupted and not available for system recovery.
This can only happen if all of the following conditions are true:
-
The client system is version 14.0 or older
-
AND the server-side system is a later release of 14.2 or higher
-
AND the old client did a Change operation.
-
AND the redo log exceeds RAM capacity and goes to the hard drive,.
.
The complete list of changes is also available in