Previous Release Notes : R2016x.R4
 
R2016x.R4
 
What's New
Other Improvements
Upgrade notes
Major fixed issues
Other fixed issues
What's New
Performance
• A lot of work has been done to ensure constant indexing time, whatever the volume of documents inside the index, for jobs with small amounts of documents.
Before this release, it was possible to observe latency peeks on jobs of similar size. You should expect homogeneous performance with this release.
Simplified CloudView upgrade
Since R2016x.R3, the upgrade between two release Refresh has been facilitated. There are only a few steps to launch the update:
• Unzip the new Refresh to a directory that will become your new INSTALLDIR
• Run INSTALLDIR/check-deps.(sh|bat)
• Change the NGINSTALLDIR variable in DATADIR/bin/ngstart.env to match your new one.
• Restart CloudView.
At restart, actions will be triggered to update in the DATADIR files that come from the INSTALLDIR (new version of cvinit.sh, new version of plugins, etc.). No change is done on config or indexed data.
Note: You can follow version migration actions in logs.
Consolidation Server
• The Consolidation Server now safely supports the partial update of documents. Relying on the Consolidation Server cache, it is possible to repush partial information only. The full document will be rebuilt before going through the transformation processor. Beware, if you build link directives on the connector side, you'll always need to repush these directives with your partialUpdate order. Only metas and parts can be used in the partial update mechanisms. 
Other Improvements
Connectors
• Add Container handling to the XML connector to be able to scan XML file within archives (#26983).
• A PushAPIFilter to convert document is available on Connector Server side (#26969). This converter should be used when pushing documents with binary parts inside a Consolidation Server. It reduces the amount of data stored inside the Consolidation Server and allows having access to meta extracted from parts in the aggregation process.
• Web Crawler: When misconfiguring the proxy, information about proxy timeout are provided in log (#27120)
Consolidation Server
• Impact detection should not aggregate the same document multiple times (#26784).
In case of large job or large impact detection, the Consolidation Server now ensures full deduplication of impacted documents.
• Payload is not mounted anymore on heap when receiving a delete order and when no delete transformation processors are defined (#26859). Clear and delete operations are processed far more quickly.
• Change of consolidation jobs tuning can be done without stopping current running jobs (#26865)
• Generate a graph synthesis of all objects and relationship stored in the Consolidation Server with cvconsole (#26871).
This feature can be useful for debugging purpose as it provides an overall view of all documents and relationships inside the Consolidation Server storage. For details, see the Consolidation Server Guide.
7809b6dbeaff02796663d5bd0f7f669b
• Full compact of Consolidation Server storage is available from cvconsole (#27038).
Core
• Default names of new Mashup UI roles deployed on slave are correctly incremented (#26802)
• Allow 'off' or 'fatal' values for logging level config (#26767)
• Improvement of Java stacks logging (#26849)
• Enumerations of PAPI-synced entries are streamed server side (#27058). They should not lead to OOM anymore.
• For Windows Server release, updating VCRedist library is made at instance installation time, not at setup time (#26796). Version level detection has also been improved.
Indexing
• Improvements on log level and message labels to reduce verbosity of indexing process (#27176)
• A new document processor checking if chunks are UTF8 valid is available (#26246)
• Hundred thousand columns in item store have a lower memory fingerprint (#26807)
• Smarter classifier to maximize density of item store at compact time (#24877)
Administration tools
• Command "cvinit.sh restart" launches the product even if it is not currently running (#27156). It was not the case in previous releases.
Convert
• Date default format to be used to interpret date cells in Excel files can be configured in Convert configuration file (#26782)
Semantic
• Support of BasisTech's open-source extra languages (#27270). It is possible, when having purchased the extended languages feature, to have access to additional language resources.
To support these languages, an additional component must be installed on your CloudView server. See the "Supported Languages V6R2016x" product datasheet for more information about extra supported languages (flagged by Ext+ ) and their add-on installation.
CAUTION: no warning is raised in this release if the component is not installed and silent fallback is made to the default language.
• Introduce guardrails into linguistic output building (#27248). Irrelevant semantic forms (mostly related-terms and phonetic forms) are dropped, therefore the amount of dictionary data is lower and the overall dictionary building performance is better.
• Allow non trivial semantic pipeline for snippetting (#27126). You can use a specific analysis pipeline to highlight correctly matched forms inside a snippet. This can be useful when having advanced tokenization configuration.
• Synonyms expansion supports plural. If a synonym rule X → Y has been defined, when searching for Xs, you also search for Y  (#14194)
• Standard Japanese Tokenizer handles patternOverrides option (#27447)
Security
• In a Security source dispatcher, it is possible to use authorization security sources only (#24398). The authentication source is no longer required.
Monitoring Console
• New graph information on Memory and CPU for all CloudView processes (for Linux release) (#27292 & #27051).
A detailed memory usage graph is also available for each CloudView process.
fbb3908c10223bd2c96d6412a615401d
• The integration of the latest JRDS version fixed the message that occured when the monitoringIntervals parameter is changed for a probe (#27096).
This probe will not collect any more until the corresponding probename.rrd file is removed.
Upgrade notes
CloudView R2016x, R2016x R1, R2016x R2 and R2016x R3 configurations are fully compatible with CloudView R2016x R4.
Warning: Due to a minor bug fix, the order in which the "before start" and "after end" optional date category facets are displayed, when sorting on date, is no longer the same.
You now see the:
• two first values when sorting by date in descending order,
• and the two last values when sorting by date in ascending order.
 
Major fixed issues
Ticket
Title
Existing since
Component
#27205
Encoding is ISO and should be UTF8 in pivot table controller
V6R2016x.R3
360/Mashup UI & API
#27343
BuildUrl fails when used with anchor
Prehistoric defect
360/Mashup UI & API
#26805
Error when installing CloudView 2016x R3 on Windows 2012R2 standard.
V6R2016x.R3
360/Plugins
#26797
Document cache compression scheme is ignored when compacting
V6R2014x
Indexing
#27284
Japanese characters are badly displayed in the test part of Analysis pipeline
V6R2016x.R3
Administration Console
#27222
Unable to access the Business Console when HTTPS is activated for Mashup Builder
Prehistoric defect
Business Console
#27168
Stats for CSV connector scan on Administration Console are not giving the number of documents pushed
Prehistoric defect
Connectors / Connector CSV
#27201
Custom Directives are ignored by the Replay connector
V6R2016x.R3
Connectors / Connector Replay
#27059
XML connector pushObjects counter stays at zero
V6R2016x.R1
Connectors / Connector XML
#26868
For heavy linked graph, consolidation heap grows far too much
V6R2015x.SP4
Consolidation
#26872
Resize of mmap file leads to crash on Windows
V6R2015x.SP4
Consolidation
#27134
Consolidation storage status grows indefinitely
V6R2015x.SP4
Consolidation
#27189
Counter for consolidation-impacted documents shows twice the amount of expected documents
V6R2015x.SP4
Consolidation
#26107
Avoid NPE while building dictionary with corrupted streams
V6R2015x.SP4
Generic
#26965
Index corruption after reboot on Windows
V6R2016x.R3
Generic
#26797
Document cache compression scheme is ignored when compacting
V6R2014x
Indexing
#26799
All build groups using a custom data model will use RowOrientedStore rather than ItemStore
V6R2016x
Indexing
#26908
Possible deadlock in StorageHttpClient.sendGET
V6R2015x.SP4
Indexing
#27099
VSD file indexing never ending
V6R2016x.R3
Indexing
#27372
NPE in ValueFieldCompacter.addSlot(ValueFieldCompacter.java:25)
V6R2016x
Indexing
#26941
Error when installing CloudView 2016x R3 on windows 2012R2 standard
V6R2016x.R3
Installer
#26993
[APP-INSTALLER] Wrong context root for application automatically created
V6R2016x.R3
Installer
#26941
Error when installing CloudView 2016x R3 on windows 2012R2 standard.
V6R2016x.R3
Installer
#23259
Custom term scoring algorithm ignored when "indexed pattern search" is activated
V6R2014x.SP3
Search
#26776
Better reporting of timeout in wildcard expansion logs
V6R2015x.SP4
Search
#26812
Replica crash on win64 during stress scenario: IndexSearch::VirtualFieldLookup::getIntegerRankingInfo
V6R2016x.R3
Search
#26834
Rare crash of summarizer when using removeDuplicateSegments
V6R2015x
Search
#27044
Incorrect boundaries and number of documents for a decimal facet
V6R2016x.R2
Search
#27067
Error on count of second dimension categories in MultiDimension Facet
Prehistoric defect
Search
#27082
Range facets with exclusiveRightBracket have strange titles
V6R2016x.R3
Search
#27143
Bad expansion of queries using nested alphanum prefixes to target multiple dynamic properties in the same field
V6R2016x.R3
Search
#27194
Bug when refine is made on facet and on multiDimension Facet
Prehistoric defect
Search
#27225
crash and/or infinite loop on ConcatValuesFacet : IndexSynthesis::ConcatFacetableCategorizer::identify
V6R2016x.R3
Search
#27339
infinite loop in MOT::MultiConverter::_computeNextToken() with invalid summarizer config
Prehistoric defect
Search
#25039
All BasisTech languages should be managed
V6R2014x.SP4
Semantic
#27389
Lemmatization matching level for Synonyms
V6R2016x.R2
Semantic
Other fixed issues
Ticket
Title
Existing since
Component
#26788
Mashup builder is not accessible after product install with different admin user
Prehistoric defect
360/Mashup Builder
#26857
Data Processing Test mode incompatible with PrecomputedThumbnailsDocProc
V6R2016x
Administration Console
#26919
Custom Prefix Handler class Id lost when modifying search logic from Administration Console
V6R2016x.R3
Administration Console
#27214
When adding a custom prefix handler, only the class is displayed, not the label
V6R2016x.R3
Administration Console
#27415
Edit button of static resource suggest disabled even for managed resources
V6R2016x.R3
Administration Console
#26070
Alerting: issue with alerts store in ORACLE
V6R2015x.SP1
Alerting
#27301
Alerting: nb_entries values have not the correct count
V6R2015x.SP4
Alerting
#14144
Relevance tuning categories get highlighted in document preview
V6R2013.SP2
Business Console
#26792
getCloudviewDeploymentInternal does not work with org.apache.cxf
Prehistoric defect
Config system / APIs
#25895
Fixed lotusnotes proxy startup
V6R2016x.R2
Connectors / Connector IBM Lotus Notes
#27018
Text content of shape in Word>=2010 docx is duplicated
Prehistoric defect
Convert
#27019
Japanese xlsx conversion adds kana-text to original text
Prehistoric defect
Convert
#27131
Fix garbage collecting of old hprof files
Prehistoric defect
Core
#27190
simple confusion about country term in java code of uc4
V6R2016x.R2
Documentation
#26786
Consolidation UC1 may lead unsuspecting end-users to get very large Consolidation Server cache
V6R2016x.R2
Documentation
#26931
In "Search API Parameters Reference", doc on secured suggest should not mention "security_token" as service parameter
V6R2016x.R3
Documentation
#27376
default value for f.facet_name.max_per_level is wrong
V6R2016x.R3
Documentation
#26864
Remove "cache.pending-documents" temp files automatically if Indexing Server crashes
V6R2016x.R2
Indexing
#27061
NPE while destroying an IndexData
V6R2014 GA
Indexing
#27267
License warning is incorrect
V6R2014x
Indexing
#26804
Analyzer crash on NDocDeserialize3 with PAPI doc containing a malformed ndoc part
V6R2013x.SP3
Indexing
#27259
Investigate on replication issues spotted during 16x functional tests
V6R2016x
Indexing
#26851
Installation fails if there is a special character in the Installer path
V6R2016x.R3
Installer
#26912
HitContentTypeValidator sends wrong warn log
V6R2016x.R3
Logging
#26733
SortHitsInGroupBy are not migrated properly from 15xSP1 to 15xSP2
V6R2015x.SP2
Migration
#26295
A template prefix handler is used inside another prefix handler is not correctly handled in the product
V6R2015x.SP4
Search
#26765
maxSliceCategories is never passed to faceting query builder
V6R2016x
Search
#26790
Wrong summary parameter default value
V6R2016x.R2
Search
#26847
Search server memory leak spotted during endurance search tests
V6R2016x.R3
Search
#26917
"Discard complex queries" doesn't work for query with prefix
V6R2016x.R1
Search
#27002
Error when MINDATE or MAXDATE aggregation functions are not used in 1st position
V6R2014x.SP1
Search
#27072
Refine on a disjunctive facet prevent queries to be cached in the search API consistent cache
Prehistoric defect
Search
#27111
Paragraph separators have no position at indexing time but have one at snippet time
V6R2016x
Search
#27306
MultiDimension Facet: first character of first category path skipped
Prehistoric defect
Search
#27378
Language for preview tokenization is hard-coded to English
Prehistoric defect
Search
#27073
Refine on a disjunctive facet make the status of a Search API query go crazy
Prehistoric defect
Search
#27300
Range size=0 - Fixed Range Numerical Partitioning
V6R2016x.R3
Search
#27243
/exa/GetStacks errors while running a CVDiag
V6R2014x.SP4
Tools
#26958
dump-attribute-group-headers command gives no details about dynamic properties
V6R2016x.R1
Tools