R2014x.SP1 is a maintenance release to R2014x. For information about new features, see the R2014x Release Notes.
Improvements
Platform
Real-time alerting
Real-time alerting leverages the R2014 indexing chain changes to reduce its CPU impact in In-memory mode (no change with previous behavior when indexing task-queue is enabled).
On in-memory indexing, real-time alerting is a document processor in the analysis chain allowing multiple alerting matchers (one per analysis thread) for better performance, reliability (not hosted on gateway = SPOF) and availability.
For more details, see "Set up real-time alerts" in the Business Console User Guide.
Document Tracker
This tool for supervising data processing was introduced in R2014x. The document cache can be disabled to make sure that you are retrieving the latest version of the document from the source instead of an older version from the cache. It also provides a full view of processed documents including parts, directives, processing flag (keep, discard...) and processing errors.
For more details, see "Testing your Analysis Pipeline Behavior" in the Technical Reference Guide.
API Console
The API Console simplifies the administration and monitoring of CloudView. It has an updated design and improved usability; it lists all available API commands, has schema completion and inline documentation for quickly sending XML commands to the API.
The "Push" allows creating documents based on data model and supports dynamic properties.
For more details, see "Perform advanced operations with the API console" in the Administration Guide.
Other platform improvements
• Indexing performance improvements: less contention and reduction of the number of written files especially for transactions impacting only a few index fields
• Optimize queries with tight sequences containing a mix of frequent words and rare words by a factor 2-10 depending of the query
• Pre-computed document thumbnails (computed at indexing time instead of query time) are now implemented as a document processor in the analysis pipeline for ease of readibility and performance improvements #19780.
• New aggregation types on facets: MINDATE and MAXDATE for computing minimum and maximum value of a date expression #15139.
• Semantic Web document processor to extract microdata & microformats from html
• Dynamically modify search logic via URL parameters For more details, see "Change search logic dynamically" in the CloudView Programmer Guide.
• New monitoring probes
o All archs: DIH size
o Windows: swap in/out, file-system usage
o Solaris, AIX: file-system usage
• Licensing screen now shows current and maximum number of user allowed #20557.
• Add comments to data model properties to document its purpose #17866.
• Installed hotfixes can be listed with the command line tool "hotfix.sh list" #19740.
• Numerical range facet now display their title even for excluded refinements #16390.
• Dynamic dates used in multidim facets display their title when refined instead of timestamps #16927.
• Dynamic dates now fully support disjunctive requirements #16053.
• Order of meta in CSV output of Search API can now be controlled, see "Search API Reference" for more details #19687.
• Hit content operations such as Truncate, Summary and Highlighting can be performed on dynamic fields #17407.
• New "Send partial results in case of error" option on search logic allows retrieving some results when a query encounters error (timeout for instance) #20108.
• Optimization of tight sequences containing stopwords #18856
Mashup
• PDF export now supports "Inline" and "Attachment" "Content-Disposition" to preview or download generated PDF.
• Search queries are now logged with "error" level if something goes wrong.
Connectors
• Connector server modifications : now supports multiple scans at the same time and active scans are restarted after process restart
• BETA release of the new Java HTTP crawler connector. The crawler now generates connectors to index documents which brings support for Push API filters. Try it out and send feedback to product management!
• Push API authentication can be configured from the "Advanced" tab of all connectors #20529.
• When a connector scan fails, its status shows error instead of idle, making it easier to track issues #19590.
Documentation
Technical Reference Guide
More improvements to the Technical Reference guide:
• New Configuring Data Modelsection with better explanations of what's happening behind the scenes:
o What is generated by the data model expansion, with diagram & step-by-step example
o Indexing options for properties with diagram
o What you can do with the analysis pipeline, with a workflow diagram
• New chapter on Configuring Search Queries:
o new User Query Language (UQL) section
o EXALEAD Low-Level Query Language (ELLQL) section
• New chapter on Organizing Search Results that includes "Defining Hit Content" section
Administration Guide
More info in the CloudView Administration Guide, for maintaining a production CloudView instance:
• New: "Controlling disk space"
• New: "Control IP address binding"
• New: "Fixing Out of Memory Issues"
• Moved: "Focus on Indexing" (previously "Configuring Indexing" in Technical Reference guide)
• Moved: "Perform advanced operations with the API console" (previously in Technical Reference guide)
Upgrade notes
• Order of columns in CSV query reporting files now matches the one defined in the UI. This might change the format of CSV reports and might require some changes if you use these files in a custom process (script...) #19914.
SDC connectors cache
If you're using an SDC connector, its configuration will be automatically migrated, but you will need to manually update its cache.
Note: This procedure rebuilds the index, which will remain usable during the update.
To upgrade your SDC connector cache
1. Go to the Administration Console Home screen
2. Under Connectors, select the More Operations drop-down list for your connector, then select Clear operations.
3. Clear these options:
o SDC document cache
o SDC object graph
o source checkpoints
4. Click Run.
5. Repeat steps 2-4 for each connector.
Documentation changes
We've reorganized the CloudView Technical Reference to be more functional, with fewer chapters and less redundant info.
For a list of new sections, see the Improvements section, above, under "Documentation".
For this info
Now see
Tech Ref, "Controlling Indexing" chapter
Administration guide, "Focus on Indexing" chapter
Tech Ref, "Using the API Console" section
Administration guide, "Perform advanced operations with the API console" section
Tech Ref, "Using Semantic Processors"
Tech Ref, Semantic processors appendix
These are now divided among 2 main groups:
• "Linguistic Processors"
• "Information Extraction Processors".
Major fixed issues
Ticket
Title
Existing since
Component
#18223
hf=-1 should be authorized at MashupAPI level but MashupUI should refuse it
V6R2014 SP2
360/Mashup UI & API
#20063
Very long inverted lists cannot be decoded (was: full compact corrupt the index)
V6R2012x
Indexing
#20149
Proximity score may be not maximal for some exact sequences
pre-V6
Search
#20244
A lot of NPE in connectors logs when scanning our exhaustive file formats folders
V6R2014 SP1
Connectors / Connector File Systems
#20387
Crash of index6 using value field with ignorePresentBit=true
V6R2014 GA
Search
#20442
XSS on error pages
V6R2014 GA
360/Mashup UI & API
#20465
DIH bloom filters sometimes empty on data, sometimes on empty slots...
V6R2014 GA
Indexing
#20476
Compact policies are not migrated :/
V6R2014 GA
Migration
#20481
New NE subcategories are not translated in french interface
V6R2014x
Semantic
#20487
Matching terms are not highlighted in preview when security is enabled
V6R2014x
Search
#20506
Category, totalResult, and others uses INTEGER instead of LONG
V6R2014 SP2
360/Mashup UI & API
#20523
stamp is never updated when pushing a document again
V6R2014 GA
Connectors / Connector File Systems
#20540
Default values are not used - depends on object class
V6R2014 SP2
Config system / APIs
#20542
Impossible to create Dynamic Date Facet with UI
V6R2014x
Admin-ui
#20543
Querying my index with or without relevance=false param return different results ...
V6R2014 SP2
Search
#20552
SDC hotfix for V6R2014x is invalid
V6R2014x
Connectors / Structured Data Consolidation
#20565
codemirror breaks Mashup Builder UI with IE9
V6R2014x
360/Mashup Builder
#20574
Slice crash on getValueFromVId
V6R2013
Search
#20594
Migration fails if named entities were installed in admin-ui and not through datamodel expansion
V6R2014x
Migration
#20627
Migration to 2014x fails if search logic's GroupBy nodes contain GroupedHitsFieldValue config elements
V6R2014x
Migration
#20630
migration from 14.SP2 to 14x fails because Unknown field/object 'index6' in DeploymentInternal
V6R2014x
Migration
#20702
sliceDownAction="error" ignored on queries with hf=0
pre-V6
Search
#20727
Very small font used for preview and thumbnail for some emails (Confex issue with <font size="XX">)
pre-V6
Convert
#20742
Only the last indexbuilder config is migrated, others are trashed
V6R2014 GA
Migration
#20744
migration fails because of empty relaxAndNodesPolicy
V6R2014x
Migration
#20750
Can't migrate an instance installed with --dynamicReporting=true
V6R2014x
Migration
#20818
BufferedPushAPI does not support "updateDocument" call
