Previous Release Notes : V6R2013.SP1
 
V6R2013.SP1
 
New features and improvements in this release
Major fixed bugs
Important notices
Known bugs
CloudView V6R2013 SP1 is a minor version of CloudView
New features and improvements in this release
Administration
It is now possible to
Add and remove and build groups in the Administration Console
Configure search targets in the Administration Console
Manage plugins (upload, update, remove) in the Administration Console
Furthermore,
New numerical and date properties in data model will now be RAM-based by default
SDK
Push and Search clients for .NET framework are now integrated in CloudView
The CloudView Programmer's Guide has been expanded
Eclipse Plugin
We are proud to present our plugin for Eclipse, making development of CloudView plugins a breeze. The Eclipse plugin supports both Core and Mashup plugins. It provides wizards for creating all custom components of the CloudView stack, to package your plugins, and live deployment on running instances.
Additionally, the Eclipse plugin provides easy creation of Mashup-UI SDK projects, creating easy-to-use development environments for widgets creation.
The plugin can be installed in Eclipse using the following URL: http://eclipse.exalead.com
Mashup UI
High availability is now supported on suggests
Query
The INNERJOIN operator now uses by default the new faster algorithm introduced in V6R2013.
Major fixed bugs
Query
Very long Japanese queries (dozens of ideograms) are now correctly handled, without triggering exponential processing times
Several crashes have been fixed when using custom term scoring
Indexing
Some categories could be lost in index when they shared a common prefix
Faceting
Several advanced options (limits, explicit sorts, range titles, ...) were not correctly handled when setting them dynamically (manually or using the "FacetingQueryBuilder" Java helpers) and have been fixed
Date and geographical facets are now correctly retrieved in hits
Several miscomputations have been fixed when geographical facets cross the 180th meridian
Several sorting issues on multi-dimension facets have been fixed
Disjunctive facets are now compatible with facets in hits
Several impossible combinations are now properly checked and rejected
Crawler
It is now possible to clear the crawler on Windows
The first pages were sometimes never indexed, this has been fixed
Security
Negative security tokens are now properly filtered
Using security sources for administration consoles authentication has been fixed
Misc
Category facets now take all entries into account instead of only the first 100
Thumbnails
Support for precomputed thumbnails is now working. Notably, cases where the task queue would not empty while using precomputed thumbnails have been fixed. Nevertheless, we still recommend to use on-demand thumbnails rather than precomputed ones.
Important notices
Migrations
Migrating a data directory from any V6 release is supported
If migrating from 12x
o After migration, the query reporting graphs are not available. You can either wait for the next daily computation, or use the API console to force a new run (Manage > search > operations > generateStaticReport and generateStaticApplicationReport)
If migrating from 12x or 12x SP1
o Russian lemmatization resource has been moved to an addon. Russian lemmatization will automatically be disabled when you migrate. If you need the feature, install the add-on and re-enable it.
o The identification of facets inside Mashup-Builder has moved from facet path (Top/MyFacet) to id (myfacet). After automatic migration, make sure that all your widgets are still displaying correct data.
o Many configuration checks have been improved. In some cases where an invalid config was previously accepted, migration might fail on the new checks. In that case, fix the reported mistake in the original configuration (you don't need to apply), then restart the migration.
o Some of the Highcharts-based Mashup widgets will fail once migrated. To fix them, you need to edit their Javascript code, and replace:
o - yAxis: {
o + yAxis: [{
o title: {
o text: null
o },
o min:0
o - },
o + }],
When migrating, the monitoring probes on the detailed status of each index field have been disabled by default, as they tended to be too costly. They must be re-enabled on a per-field basis if you need them, by editing config/Reporting.xml, and inserting, within the <PerformanceMonitoringConfig> node, children nodes like <MonitoredIndexField name="text"/>.
API Compatibility
APIs are fully compatible with V6R2013.
Compared to V6R2012x, the following changes apply (introduced in V6R2013):
o Crawler forced metas and feedfetcher metas are no longer prefixed with meta: or ff:.
o "MatrixFacet" has been removed and is replaced by "MultidimensionFacet" and "Hierarchical2DFacet". Please see the CloudView Technical Reference Guide to know which one you need.
o The following deprecated parameters have been removed on the SearchCommand class (use the SearchCacheinstead):
cacheQueriesqueryCacheNbElementsqueryCacheTimeToLiveSqueryCacheTimeToIdleS
Known bugs
Under heavy build load, reattaching a detached slice can make the slice process restart.
The excluded category of a MultiDimension (or Hierarchical2D) facet is lost, and therefore cannot be canceled
When a 2D data table is built from a MultiDimensionFacet, the double aggregation values are displayed as integers
Reporting problems
Always use CVDiag to report crashes and errors. To access CVDiag from the Administration Console, go to Monitor > Reporting.