 | Summary: Fixed issues in GigaSpaces 6.0.2 (build 2002).
|
Overview
This list includes fixed issues since GigaSpaces 6.0.1 (build 1903).
See the list of new features and known issues in GigaSpaces 6.0.1.
Upgrading GigaSpaces 6.0.1? See this migration document.
Configuration, Jini Services, and Tools
- GS-3338 – Fixed: A space container couldn't be found using the Jini protocol and providing a host name. Workaround: Works when providing an IP address instead of host name.
Workaround: Provide an IP address.
- GS-3256 – gsInstance and puInstance now use bootclasspath (adding XML JARs) when starting the
Java process. It is required for 1.4 VM (or Mac 1.5 VM) when running a primary-backup topology due to a xalan parser bug fixed in later releases.
- GS-3202 – Fixed: LOOKUP_GROUPS set in setenv.bat in XAP EDG 1903 ended with XAP and not EDG.
Support case 3366
- GS-3127 – Fixed: RMI registry ran within a GSC as well as within a space container. This caused a problem when starting a RMI registry in a particular port, and enabling a space container running within a GSC bound to the same port. It caused the space deployed into the GSC not to be found.
Support case 3366
- GS-3090 – Fixed: Spaces were not added to the container, since the space was registered in the Jini Lookup Service with the IP address set as the HostName value.
Support case 3326
- GS-2263 – Fixed: Unicast locators and groups in a space were appended when an IWorker included code to find a space handler for another space.
Support case 3062
- GS-3067 – Fixed: Unicast locators were appended when multiple spaces were started in the same VM.
For more details, see Lookup Service essential guidelines section.
- GS-3055 – Fixed: The initialUnicastDiscoveryPort was missing from the Reggie/Mahalo configuration in the services.config file.
initialUnicastDiscoveryPort = 1222;
Set the LOOKUPLOCATORS system property in <GigaSpaces Root>\bin\setenv.bat/sh to match the port number you defined (in this case, pc-lab:1234). That is required if you specify an explicit unicast/locators port, otherwise the service will use the fallback 4160 port if not set explicitly.
For more details, see Reggie JavaDoc.
- GS-3002 – JSpaces.jar has been added to the service browser (lookupbrowser) classpath in order to avoid NoClassDefFound exceptions.
- GS-2914 – Fixed: Starting more than one space in a single JVM with different lookup group names/locators caused the group names/locators to be appended.
Support case 3236
- GS-2485 – Fixed: Entry serialization was inconsistent with Java serialization.
Support case 3135
GigaSpaces Management Center
- GS-3215 – Fixed: write and notify operations did not work with a clustered proxy when the Use Clustered Proxy check-box was selected.
- GS-3253 – Fixed: In the case of hundreds of clients, several views (such as connections, transactions, classes and more) in the GigaSpaces Management Center did not show data.
- GS-3092 – Fixed: A FIELD serialization exception was thrown in the GigaSpaces Management Center and the cluster wasn't displayed properly, due to the re-use of the clustered proxy when a load-balancer was used. The proxy returned a "used" clustered proxy.
Support case 3326
- GS-3074 – Fixed: The Transactions view displayed the wrong ID when performing XA transactions.
- GS-3073 – Fixed: The Transactions view showed the lease time instead of end time.
Core
- GS-2783 – Fixed: Space security didn't support Externalizable.
Support case 3215
- GS-3329 – Fixed: The broadcast mechanism ignored the routing property null value, so the broadcast operation was always routed to a single space.
- GS-3301 – Fixed: The protection level of transactional operations like abort or commit was the same as the write operation, instead of the read operation (since the data is already in the space).
- GS-3295 – Fixed: NegativeArraySizeException was thrown when performing notify replication with a new object as its template.
- GS-3263 – Fixed: An internal error occurred when processing notify events during failover in a large cluster.
- GS-3242 – Fixed: A NullPointerException was thrown after closing notifications.
Support case 3381
- GS-3233 – Fixed: The RecoveryValidator was not cleaned after the source space was restarted.
Support case 3174
- GS-3210 – Fixed: An OutOfMemoryException was thrown when many clients read large Entries from space.
- GS-3183 – Fixed: When failure occured in a primary/backup topology where several threads were performing read and take operations, it took tens of seconds for some threads to recover and continue functioning.
Support case 3360
- GS-3171 – Fixed: When working in synchronized replication mode with many clients, the replication channel between the primary and backup space was disconnected; and when attempting to reestablish the connection, the recovery mechanism was corrupted.
Support case 3349
- GS-3160 – Fixed: An ArrayIndexOutOfBoundsException was thrown when setting EventSessionConfig.setBatch(0,0).
- GS-3118 – Fixed: SQLQuery might have returned a wrong result when the query contained a routing index and was used with a partitioned schema.
Support case 3274
- GS-3117 – Fixed: A MemoryShortageException was thrown after the space was cleared, even though there was enough memory.
- GS-3085 – Fixed: Low performance when performing clear on a large quantity of Entries in a sync-replicated cluster.
Support case 2454
- GS-3083 – Fixed: A ClassCastException was thrown when performing the updateMultiple operation when the same POJO appeared in the space twice.
Support case 3316
- GS-3081 – Fixed: Multi-NIC didn't work since the java.rmi.server.hostname property was not read in several places.
- GS-3070 – Fixed: when injecting an external space filter, the enabled flag was discarded.
- GS-3031 – Fixed: XA transactions were stuck when enlisting 2 clusters and preforming space operations in both of them.
- GS-2840 – Fixed: Clearing a class in a secured space with the same user performing write operations worked, while write operations failed in the same settings.
- GS-2839 – Fixed: A clean operation could be performed in a secured space by any user that had defined roles.
- GS-2453 – Fixed: Failover and recovery problems with Service Grid and primary election.
.NET
- GS-3320 – Fixed: The .NET proxy failed when using IPreparedTemplate, and the space did not contain metadata. IPreparedTemplate now contains two copies of the template, with and without metadata.
- GS-3056 – Fixed: Partial update didn't work with arrays.
- GS-3000 – Fixed: LocalCache didn't work with clustered spaces in .NET.
- GS-2939 – Fixed: Write with the UpdateOnly modifier, local cache, and versionid was not increased when updated in the space.
- GS-2938 – Fixed: A NullPointerException was thrown when performing UpdateMultiple in a partitioned space.
Service Grid
- GS-3246 – Fixed: com.gs.home was not defined when used with Service Grid override files.
Support case 3318
OpenSpaces
- GS-3145 – When registering for notifications using a notify container which accesses a remote space: If you are working directly with a cluster member that was started in embedded mode, the value is true. Otherwise, the default value is based on the cluster schema (true for clusters with backups).
Workaround: Set the replicate-notify-template flag to true.
- GS-3071 – Fixed: annotation and method-adapter event listener schema configuration failed when setting the write-lease parameter.
