Release Note 20140603

Table of Contents

Features & Improvements

This is a summary of the new features and improvements introduced in this release:

Console: Redesigned Top Bar

Completely redesigned the Console Top Bar and reorganized the content in 3 categories: Feedback, Help, and Profile (User name).

Under Help, these are the shortcuts available:

  • My support cases
  • Support chat
  • Documentation
  • Changelog (Release Notes)
  • Service status
  • Create support request

The User menu has shortcuts for the User Profile and Logout.

Console: Redesigned Top Bar

Console: Treasure Agent Monitoring Service (Public Beta)

We are excited to announce the Public release of the Treasure Agent Monitoring Service. By adding a simple to deploy plugin, users can track the status of their td-agent (Treasure Agent daemon) on the Console and monitor its performance.

We have also added notifications to allow the user to get an email when any of their monitored td-agent instances goes down for more than 5 minutes.

For more information, please visit the Treasure Agent Monitoring Service (TAMS) documentation page.

Console: Treasure Agent Monitoring Service (Public Beta) Console: Treasure Agent Monitoring Service (Public Beta) 2

Console: Formatted/Raw Query for Schedules

The ‘Query’ field view for Scheduled query can also be visualized in Raw or Formatted format as already supported by on-off Jobs.

Console: Formatted/Raw Query for Schedules

SDKs: Released iOS v0.0.3

This is the first beta release of the iOS SDK.

For more information, please visit the Treasure Data iOS SDK documentation page.

Client Libraries: Released Java Client Library v0.4.2

The new release of the Java client library mainly adds support for SSL communication with the default treasuredata.com API domain.

Please refer to the Github Changelog.

JDBC Driver: Released JDBC v0.2.9

This new version of the driver builds on top of the newly released Java Client Library (see above) adding supports for SSL communication with the Treasure Data APIs.

Please refer to the Github Changelog.


Bug Fixes

These are the most important Bug Fixes made in this release:

Console: Edit as New Query

  • [Problem]
    “Edit as New Query” for a job with an unsaved output target containing a password sets the password wrongly when saving the job as Schedule.
    [Solution]
    When saving to a Schedule the original job unnamed output target’s password was not retrieved from the database.

Console: Manual Schedule Runs

  • [Problem]
    Date-time used when manually triggering a job for a Schedule query does not match the date-time selected.
    [Solution]
    The problem was due to the date-time picker in the Schedule view.
    The date-time picker has been changed to a stock AngularJS-bootstrapped one and fixed.

Console: CPU Core Utilization

  • [Problem]
    The CPU core utilization chart may show non-zero utilization even when the system is not used.
    [Solution]
    Since 0 CPU utilization data points are not logged, non-zero utilization points were connected by a line inducing the observer into thinking the utilization was non-zero during the times in between the non-zero data points.
    Fixed the visualization by inserting zeros for all points where the CPU utilization is expected to be 0.

APIs: Schema Auto-Detection on Import

  • [Problem]
    Schema auto-detection can fail to detect sparse columns (columns with elements in a very small subset of the records).
    [Solution]
    Schema auto-detection is performed only on the records stored in the tail but only a small portion (40) of the records imported is stored in the tail. The tail may happen to contain none of the records with elements from the sparse columns.
    Schema auto-detection is extended to be performed on all records imported.

APIs: Manual Schedule Runs and Timezone Changes

  • [Problem]
    Manual run requests for Schedules whose timezone of reference has been modified fail with a 500 (Internal Server) error.
    [Solution]
    This was caused by a mishandling of long-form (Tokyo, Los Angeles, etc…) timezone specifiers.
    Now all timezones are referred back to their unique numeric identifier for handling consistency.

Backend: Handling of Hive’s Semantic Error

  • [Problem]
    Hive queries retries (12 times, up to 1 hour) on Semantic error exceptions.
    [Solution]
    Semantic error exceptions are not included in the list of unrecoverable error codes for Hive.
    Unrecoverable errors are those deterministic errors, errors whose result won’t change if retried.

Worker: Data Exporting

  • [Problem]
    Exporting data from Treasure Data to AWS S3 can cause Null Pointer Exceptions.
    [Solution]
    This was due to having keys with nil values in the parameters hash map when constructing an error message.
    Replacing the value of nil keys with the string “null”.


Last modified: Jul 11 2014 00:40:43 UTC

If this article is incorrect or outdated, or omits critical information, please let us know. For all other issues, please see our support channels.