================================================================================ Oracle Big Data Discovery (1.1.3) GENERAL (1.1.3) - Software Release Date: May 2016 - The release notes and related product documentation may have been updated since the release date. The latest release notes and documentation are available through the Oracle Big Data Discovery documentation page on the Oracle Help Center at https://docs.oracle.com/cd/E64107_01/index.htm or by contacting Oracle Technical Support for Big Data Discovery. INSTALLATION AND UPGRADE (1.1.3) - Refer to the "Oracle Big Data Discovery Installation and Deployment Guide" for installation information. - See the "Oracle Big Data Discovery Upgrade Guide" for information on upgrading to this release. BUG FIXES (1.1.3) Installation: - EADMIN-1424: Fixes a problem where the installer didn't verify the full version number of the JDK. - EADMIN-1641: Fixes a problem installing BDD on HDP required two Hadoop configuration files to be edited. This step is now performed automatically by the installer. Administration: - EADMIN-1401: Fixes a problem that occurred on machines whose system language was set to a non-English language in which the bdd-admin status command would report that components were running even though the start command failed to start them. - EADMIN-1661: Fixes an issue that caused the bdd-admin script to fail when the INSTALL_TYPE property in bdd.conf was set to BDA. Data Processing: -EDPROC-3669: Previously, Data Processing startup would fail when the HTTP endpoint for Yarn daemons was configured to be HTTPS_ONLY. This problem has been fixed. Upgrade: - EADMIN-1663: Fixes an issue where the upgrade script would disable the Hive Table Detector in clusters that had it enabled. Documentation: - EDOC-2111: WebLogic Server was removed from the list of packages to download before upgrading BDD, as this is not required. KNOWN ISSUES (1.1.3) Installation: - EADMIN-1813: Installing on BDA 4.3 and 4.4 can fail because those versions of BDA only allow BDD 1.1.1 to be installed. A workaround for this is to open the file /opt/oracle/BDAMammoth/puppet/modules/bdd/templates/bdd_prereq_script.erb and remove the extra space from the following line: /bin/sed -i 's/bdd_min_ver $/bdd_min_ver/' So that it reads: /bin/sed -i 's/bdd_min_ver$/bdd_min_ver/' Administration: - EADMIN-1401: When running on a machine whose system language is set to a non-English language, the bdd-admin status command can report that a component is running even though the bdd-admin start command failed to start that component. The workaround is to set the system language to English. ================================================================================ Oracle Big Data Discovery (1.1.1 Revision A) GENERAL (1.1.1) - Release Notes Updated: December 2015 - This documentation-only update reflects the availability of Version 13c Release 1 of the Enterprise Manager Plug-In for Big Data Discovery. The plug-in is intended for Enterprise Manager Cloud Control 13c. The plug-in is available as a separately-downloadable module, as described in the "Enterprise Manager Plug-In for Big Data Discovery Installation Guide". BUG FIXES (1.1.1) Documentation: - EDOC-1286: "Enterprise Manager Plug-In for Big Data Discovery Installation Guide" has been updated for Version 13c Release 1 of the Enterprise Manager plug-in. - EDOC-1287: "Administrator's Guide" has been updated to reflect the features of Version 13c Release 1 of the Enterprise Manager plug-in. - EDOC-1328: "Installation and Deployment Guide" references the "Oracle Big Data Appliance Owner's Guide Release 4 (4.3)" for BDD installations on Oracle Big Data Appliance (BDA) or Oracle Public Cloud (OPC). - EDOC-1373: "Installation and Deployment Guide" clarifies that connections of a single BDD instance to multiple Hadoop clusters (for example, multiple BDAs) are not supported. In other words, BDD can only connect to one Hadoop cluster. - EDOC-1389: "Upgrade Guide" now describes how to troubleshoot the problem of the Data Processing component failing to start. - EDOC-1401: "Installation and Deployment Guide" adds that the Linux user performing the BDD installation needs permissions to create the BDD directory specified by the ORACLE_HOME property in the bdd.conf file. - EDOC-1462: "Installation and Deployment Guide" adds the download links for the supported JDK versions. - EDOC-1483: "Data Exploration and Analysis Guide" provides more information on escaping special characters in regular expressions in Transform functions. - EDOC-1486: "Installation and Deployment Guide" updates information for the required Perl modules. - EDOC-1487: "Data Exploration and Analysis Guide" clarifies that page permissions in Studio are determined by project permissions. - EDOC-1531: "Data Exploration and Analysis Guide" documents the limitations on the IFrame component. KNOWN ISSUES (1.1.1) Administration: - EADMIN-1401: When running on a machine whose system language is set to a non-English language, the bdd-admin status command can report that a component is running even though the bdd-admin start command failed to start that component. The workaround is to set the system language to English. ================================================================================ Oracle Big Data Discovery GENERAL (1.1.1) - Software Release Date: October 2015 - The release notes and related product documentation may have been updated since the release date. The latest release notes and documentation are available through the Oracle Big Data Discovery documentation page on the Oracle Help Center at https://docs.oracle.com/cd/E64107_01/index.htm or by contacting Oracle Technical Support for Big Data Discovery. INSTALLATION AND UPGRADE (1.1.1) - Refer to the "Oracle Big Data Discovery Installation and Deployment Guide" for installation information. - See the "Oracle Big Data Discovery Upgrade Guide" for information on upgrading to this release. BUG FIXES (1.1.1) Administration: - EADMIN-1182: Previously, backup would fail if Kerberos was enabled. This problem has been fixed. - EADMIN-1212: Previously, when Oracle Database was used for Studio, you would see redundant data in Studio's Configuration page after restoration. This problem has been fixed. - EADMIN-1258: Previously, collecting zookeeper logs by running the bdd-admin script with the get-logs command would fail on HDP clusters that had Kerberos enabled. This problem has been fixed. - EADMIN-1265: Previously, Studio sandbox files were not backed up, which caused the creation of data sets from transformed data to fail after BDD had been restored. This problem has been fixed. - EADMIN-1291: Previously, when you tried to obtain Data Processing logs by running the bdd-admin script with the get-logs command, the logs would not be collected if Studio wasn't installed on Hadoop nodes. This problem has been fixed. Data Processing: - EDPROC-2558: Previously, if you created a data set (with auto-enrichment enabled) using a file or a Hive table that has columns that resulted from auto-enrichment, then applying any Transform operation would fail. This problem has been fixed. Dgraph: - ESERVER-1403: Previously, the Dgraph might shut itself down if it disconnected from ZooKeeper and reconnected to it during an ingest or data update. Now, the Dgraph handles this situation gracefully. - ESERVER-2135: Previously, the Dgraph could operate for a few seconds as if it were successfully communicating with ZooKeeper when it had actually lost the connection to ZooKeeper. - ESERVER-2173: Previously, the Dgraph might be unable to re-promote to leader if it lost its leadership status as a result of losing its connection with ZooKeeper until the Dgraph is restarted. This is no longer the case. - ESERVER-2277: Previously, the Dgraph was not deleting spelling dictionaries on disk that were obsolete. This behavior could cause the Dgraph to start up slower than normal. Now, the Dgraph deletes such files. - ESERVER-2360: Previously, the behavior of the Dgraph upon encountering an EQL keyword constant (like MINUTE or YEAR) in certain contexts was not well-defined and could potentially include inappropriate error messages or even a Dgraph crash. The Dgraph now signals a well-defined error in this case. Studio: - ESTUDIO-7234: Fixed an issue where, in some instances, when you configure or process incremental updates, the Data Set is incorrectly labeled as "Sampled". - ESTUDIO-7282: Fixed an issue in Box Plot where attribute selections could not be changed after unchecking "Auto Binning". - ESTUDIO-7298: Fixed an issue with the Histogram Plot where you were unable to set a custom number of bins. - ESTUDIO-7489: Fixed an issue with the Custom Visualization Component where you could lose your JavaScript code if you canceled out of the JavaScript editing modal and clicked save without modifying your JavaScript. - ESTUDIO-7501: Fixed an issue with the Custom Visualization Component where external CSS resources were not getting requested. - ESTUDIO-7512: Fixed an issue where in Transform, attempting to replace a committed transform script by loading a new script would result in appending to your script instead of replacing your script. - ESTUDIO-7546: Fixed an issue where using JDBC import to import data from an Oracle database would in some cases result in a "FROM keyword not where expected" error. - ESTUDIO-7635: Fixed an issue where chart configuration would fail in some cases when two chart components were on the same page. - ESTUDIO-7707: Fixed an issue in Transform where some functions in the function list were missing descriptions. - ESTUDIO-7923: Fixed an issue in the Chart component where some bubbles outside of the chart range would display and conflict with the legend. KNOWN ISSUES (1.1.1) Administration: - EADMIN-1266: The EM Plug-in currently only supports the English language, even though the EMCC platform supports many languages. Studio: - ESTUDIO-1978: In some cases, very small double precision numbers (i.e. 1.0E-324) may be incorrectly indexed and displayed as ¿.0E-324. This may cause errors in transform when attempting to apply transformations to numbers of that type. - ESTUDIO-3334: The Parallel Coordinates Plot does not support multi-assign attributes. - ESTUDIO-5601: Timeline: Timeline render fails when a data set has only one date/time multi-assign attribute. - ESTUDIO-6125: Health Check: The result of health check doesn't have a timeout mechanism. - ESTUDIO-6410: When you navigate to the Transform page of a newly created project, the sort options list is initially missing the "By custom column order" choice. Refreshing the page fixes the problem. - ESTUDIO-7706: During an upgrade from release 1.1.0 to release 1.1.1, sometimes Studio will to fail to start properly. Restarting all of BDD solves the problem. - ESTUDIO-7946: The thematic map component hangs if you perform a marquee selection that contains only states that lack data. - ESTUDIO-8187: The Change Locale dialog fails with a "Portlet is temporarily unavailable" message for non-Administrative users until a user with Admin privileges performs the operation. - ESTUDIO-8344: Double clicking the Results List component "Record detail" link opens two modals instead of only opening one. ============================================================================= Oracle Big Data Discovery GENERAL (1.1.0 Revision A) * Release Note Update: August 6, 2015. * Revision A consists of an update to this release note file. The list of release note entries in the sections below has been corrected. The software has not changed since the 1.1.0 release. ================================================================================ Oracle Big Data Discovery GENERAL (1.1.0) - Software Release Date: August 2015 - The release notes and related product documentation may have been updated since the release date. The latest release notes and documentation are available through the Oracle Big Data Discovery documentation page on the Oracle Help Center at https://docs.oracle.com/cd/E64107_01/index.htm or by contacting Oracle Technical Support for Big Data Discovery. INSTALLATION AND UPGRADE (1.1.0) - Refer to the "Oracle Big Data Discovery Installation and Deployment Guide" for installation information. - See the "Oracle Big Data Discovery Upgrade Guide" for information on upgrading to this release. BUG FIXES (1.1.0) Installation: - EADMIN-493: Using symlinks in the installer's configuration file will cause deployment to fail, as the installer resolves all symlinks to their full paths. To avoid this, only use absolute paths in the configuration file. - EADMIN-625: Previously, if the $JAVA_HOME environment variable was not set to the correct location of Java, the installer could fail with this incorrect error message when checking the ZooKeepers: "Fail! COORDINATOR_INDEX already exists in ZooKeeper, please use another value." Now, if the Zookeeper-client execution fails, it will report: "Fail! Error executing zookeeper-client on ${zkServer}". Data Processing: - EDPROC-1480: Previously, after committing a transform that includes the Address GeoTagger, the Post Code attribute that gets created was mistakenly detected as a number type and, when shown in Studio, was formatted as a number, with leading zeros trimmed and commas placed in the thousands column. - EDPROC-1623: Previously, Spark would fail to clean up cache/lock files in the /tmp directory on Spark Worker nodes. This could result in /tmp getting filled up. This problem no longer exists with the move to the Spark-on-YARN architecture of this release. - EDPROC-1640: This fixes a problem in which newly added and enriched attribute values were all empty after committing a Tranform. The problem occurred for String columns that contained less than 100% geocode types and Data Enrichment was run. Dgraph: - ESERVER-1066: If there is a severe network problem and the Dgraph disconnects from ZooKeeper, a follower Dgraph node may segfault as a result of catching up while disconnected from ZooKeeper. This problem has been fixed. - ESERVER-1481: Previously, if there was a severe network problem and views (semantic entities) were being updated, queries referencing views could return incorrect results. In the current release, re-try logic has been added to ZooKeeper operations. Studio: - ESTUDIO-3285: Fixed a problem where new multi-assign attributes that were created as the result of a transform operation were set to Single Assign. - ESTUDIO-3577: Previously, when bringing up the custom Transformation Editor from the "Add attribute" menu, the "Use refinement state as a conditional statement" was greyed out. It is now available and selectable in this situation. - ESTUDIO-3771: Fixed a problem where the aggregation functions for attributes in linked data sets reverted to default values when the links changed. - ESTUDIO-3842: Fixed a problem where a Histogram Plot displayed incorrect data when the histogram type was set to cumulative. - ESTUDIO-3796: Fixed a problem where attributes marked as Favorites were not displayed in the first columns of the Transform page after logging out and logging back in again. - ESTUDIO-3874: Previously, the "Changeable" option must be checked when configuring the "Password Policies" page in the Control Panel. The solution is that the property "Changeable" in "Password Policy" is set to "true" when creating a new user. - ESTUDIO-3944: Fixed a problem where moving a column on the Transform grid required you to refresh the page before you could move another column. - ESTUDIO-3964: Fixed a problem where errors could occur when transforming different data sets in separate tabs of the same Web browser. - ESTUDIO-3968: Fixed a problem where the left and right scroll bar arrows did not work in the Pivot Table if you were using Internet Explorer versions 10 or 11. - ESTUDIO-3997: Fixed a problem where charts were not included in snapshots if you used Internet Explorer versions 10 or 11. - ESTUDIO-4081: Resolved a problem where the "Title" and "pt" labels were missing from the "Color" format option in the Chart component. - ESTUDIO-4144: Fixed a problem that when refining on a Date Time attribute used as the Trellis attribute only the first trellis displayed. - ESTUDIO-4222: Resolved a problem where custom views that contained a pre-defined metric could not be exported. - ESTUDIO-3862: Fixed a problem where the label on X-axis was missing after switching to Trellis mode. - ESTUDIO-3931: Fixed a problem in the Chart component where the Y-Axis labels in trellis were truncated if the values were too long. - ESTUDIO-4204: Fixed a problem where an Unpublished view could be used to generate the default map layer. - ESTUDIO-4504: Fixes a problem where a Bookmark did not capture Chart configuration. - ESTUDIO-5274: Fixed a problem where the attribute selection of Chart component was not working correctly with Internet Explorer 10. - ESTUDIO-6328: Fix a problem that Snapshot of charts skewed in Firefox. - ESTUDIO-4382: Previously, scatter chart will crash if user select multi-assigned attributes. The solution is to split multi-assigned attributes by individual values before the aggregation. - ESTUDIO-4860: Fixes a problem that area chart sometimes shows as a line chart. - ESTUDIO-5240: Fixes a problem that stacked vertical bar chart shows only one metric(bar) while there are two metrics. - ESTUDIO-5652: Fixes a problem that chart will not show when using the same metric for a multi-assign dimension. - ESTUDIO-6200: Fixes a problem that the multi-select queue will display multiple range filters, but only apply one as a refinement. - ESTUDIO-6404: Fixed a problem that in Explore where the values on the X-axis overlapped in the Scratchpad. - ESTUDIO-6601: Previously, the attribute name in an uploaded file was case sensitive. Now the attribute name is case insensitive. - ESTUDIO-7109: Fixed a problem where Geospatial range filtering, text search, and auto pan did not display properly in the Map component. - ESTUDIO-7243: Fixed a problem in the Chart component where the "records with values" aggregation was not shown in the Legend. - ESTUDIO-7259: Fixed a problem in Transform where rearranging attributes made them invisible in Explore. - ESTUDIO-7108: Fixed a problem in Project Settings where a User is taken to the Catalog after changing locale value on the Project Settings page. - ESTUDIO-6430: Fixed a problem in Explore where a newly added dataset or view was not included in some components. - ESTUDIO-6194: Fixed a problem in Explore where Datetime charts did not cascade in Scratchpad. - ESTUDIO-5567: Fixed a problem in Explore where the Attribute filter bar used a key value instead of the display name to filter. - ESTUDIO-5565: Fixed a problem in Data Set Details footer: when the full data set is loaded, the footer should display "x Records" (not "x of x Records"). - ESTUDIO-5417: Fixed a problem in Explore where switching the date increment did not take effect in the Explore detail window. - ESTUDIO-5147: Fixed a problem in Explore where switching visualizations in Scratchpad preserved alternative visualizations. - ESTUDIO-4559: Fixed a problem in Data Views where pre defined metrics did not show up correctly if they have special characters. - ESTUDIO-4439: Fixed a problem in the Histogram Plot component where it did not inherit the value formatting defined in Data Views. - ESTUDIO-4357: Fixed a problem in Explore where "Save to discover page" did not work if Studio was set to non-English locales. - ESTUDIO-4259: Fixed a problem in Explore where the page focused on old data sets even after new data sets were added. - ESTUDIO-4167: Fixed a problem in the Map and Results Table components where they could show NAN and INF for numeric columns instead of empty string. - ESTUDIO-4135: Fixed a problem in the Results List component where data loaded incorrectly if datasets were linked. - ESTUDIO-3893: Fixed a problem in searching the Catalog where data sets did not always return in search results. - ESTUDIO-3920: Fixed a problem in Available Refinements where the filter results were incorrect on the view page. - ESTUDIO-3938: Fixed a problem in the Search box where search value were not highlighted. - ESTUDIO-4003: Fixed a problem in Export where deleted data sets could be selected in the Bulk Export. - ESTUDIO-4117: Fixed a problem in Export where it was easy to mistakenly cancel export to desktop. - ESTUDIO-4190: Fixed a problem in Summarization Bar where the Left side-bar was still available to edit even when a modal dialog was opened. - ESTUDIO-4224: Fixed a problem on the Roles page where the Edit panel cannot open if there was line wrapping line in the Description field. - ESTUDIO-4245: Fixes a problem in Export where alert message for refinements did not display with linked views. - ESTUDIO-4348: Fixed a problem on the Control Panel > Credentials page where a user could not log in if their password was not set to Changeable. - ESTUDIO-4423: Fixed a problem in Export where you were unable to check hive table name after the hive service stopped and restarted. - ESTUDIO-4456: Fixed a problem in Export where an export would fail if a column name started with an underscore '_'. - ESTUDIO-4474: Fixed a problem in Password Policies page where the Change Required Password Policy only worked the first time. - ESTUDIO-4808: Fixed a problem in Export where checking for Hive table name overlaps occurred immediately. - ESTUDIO-5341: Fixed a problem in Password Policies page where the Changeable option be disabled for LDAP users. - ESTUDIO-5479: Fixed a problem in Results Table component where extra commas in the data caused an issue during export. If you created a new data set, based on the exported data, the data set would contain an empty attribute. - ESTUDIO-5744: Fixed a problem in Export where the export operation failed if the Hive table name already exists. - ESTUDIO-5797: Fixed a problem in Export where the specified file name was not used during export. - ESTUDIO-5817: Fixed a problem in Export where a columns were omitted from export if the "Interchangeable Column Sets" option was used. - ESTUDIO-5818: Fixed a problem during Export in the Results Table. The component could have more columns exported than were selected in a custom view. - ESTUDIO-6062: Fixed a problem in Users page where deactivated users were still available in the users list of a project or data set. - ESTUDIO-6691: Fixed a problem in Transform where the "Create a Data Set" dialog did not permit spaces in data set names. - ESTUDIO-4530: Fixed a problem in the Box Plot component to reduce the amount of space between plots to render more plots in the view area. - ESTUDIO-5298: Now using unique breadcrumb labels for data set and project authors. - ESTUDIO-5554: Wildcard tokens are now added before and after search terms in Catalog. - ESTUDIO-6258: Improved logic for axis scaling so that we no longer show duplicate labels. - ESTUDIO-6611: Modified the Box Plot component so that refinement only requires a single click. - ESTUDIO-6644: Modified the Box Plot component to support multi-assign attributes. - ESTUDIO-6995: The Box Plot component now shows min and max outliers for all plots. - ESTUDIO-4005: Fixed an issue where the label for the bdd.sparkExecutorMemory indicated that it defaulted to 10g when in fact the default is 48g. - ESTUDIO-4139: Fixed an issue where you were required to click the Save button twice on the "Create Data Set" dialog. - ESTUDIO-4151: Fixed a problem with the Data Set Relationships where the diagram of data set relationships would sometimes be jumbled. - ESTUDIO-4171: Fixed an issue where the Thematic Map component would not zoom in and out from the center. This could cause the map to zoom out of view of the display area. - ESTUDIO-4247: Fixed an issue where some transform editors would lose state when changing the navigation state. - ESTUDIO-4260: Fixed an issue where the linked views that were created as the result of linking data sets would remain after the data set link was removed. - ESTUDIO-4934: Fixed an issue with refinement behavior of multi-AND attributes. The refinement behavior allowed you to select refinements that would lead to no results. - ESTUDIO-5296: Fixed an issue where the group editor in transform would not correctly handle values that contained single or double quotes. KNOWN ISSUES (1.1.0) Data Processing: - EDPROC-2558: If you create a dataset (with auto-enrichment enabled) using a file or a Hive table that contains columns that resulted from auto-enrichment, then applying any Transform operation will fail. A workaround is to change the names of the auto-enrichment output columns in the file or Hive table before creating the data set. Dgraph: - ESERVER-1188: PropertyValueSearch does not take join filters into account. - ESERVER-1199: If the leader Dgraph is shut down during an ingest, follower Dgraphs may not be able to promote afterwards. The formerly leader Dgraph will be able to promote, though, if it is started back up. - ESERVER-1623 and ESERVER-2069: Very rarely, if ZooKeeper is down, after trying to reconnect for a period, Dgraph may crash or cease all reconnection attempts. - ESERVER-1403: If the Dgraph loses its ZooKeeper connection during ingest or data update, the Dgraph may shut itself down after ZooKeeper comes back up. - ESERVER-1495: If you turn off ZooKeeper for a time longer than the session time-out, it will take until another session time-out's length of time before a new leader can be selected. Studio: - ESTUDIO-1978: In some cases, very small double precision numbers (i.e. 1.0E-324) may be incorrectly indexed and displayed as ¿.0E-324. This may cause errors in transform when attempting to apply transformations to numbers of that type. - ESTUDIO-3334: The Parallel Coordinates Plot does not support multi-assign attributes. - ESTUDIO-5601: Timeline: Timeline render fails when a data set has only one date/time multi-assign attribute. - ESTUDIO-6125: Health Check: The result of health check doesn't have a timeout mechanism. - ESTUDIO-7298: The Box Plot does not support a custom number of bins. - ESTUDIO-7282: In Box Plot, attribute selections cannot be changed after unchecking "Auto Binning". - ESTUDIO-6410: When you navigate to the Transform page of a newly created project, the sort options list is initially missing the "By custom column order" choice. Refreshing the page fixes the problem. - ESTUDIO-7234: In some instances, when you configure or process incremental updates, the Data Set is incorrectly labeled as "Sampled". =============================================================================