3.1.2

Includes significant refactoring, JUnit changes, and Bash scripts.

Changes to connection profiles:

• Use ${varname} for environment variable interpolation.
• Add property loader.metrics_folder

Bash scripts in bin directory:

• sndml-setup - define bash functions
• load-fg - load foreground
• load-bg - load background
• load-table - load a single table
• generate-create - generate SQL for CREATE TABLE
• delete-old-logs

Bash scripts and JUnit tests assume all files are in the configs directory (previously this was named profiles).

3.1.1

3.1.1 includes significant restructuring of JUnit tests.

Also included is a new columns feature which restricts the columns that get loaded. Here is an example:

tables:
- {name: incident, columns: number state assignment_group assigned_to short_description}

The columns argument can be a comma delimited or a space delimited list of field names. (A space delimited list is easier with YAML because you do not need to put it in quotes.)

sys_id, sys_created_on and sys_updated_on are always loaded, regardless of whether or not they are included in the list.

The columns setting uses sysparm_fields to reduce the amount of data retrieved from ServiceNow. In theory, it should be possible to load large tables faster by using a short columns list and a large pagesize. It does not affect table creation. The database table will be created with all available columns, regardless of whether or not they are loaded.

3.1.0

3.1.0 includes significant changes

• The JDK has been upgraded from 1.7 to 1.8
• Open source drivers are now used for Oracle and MS SQL JDBC
• As a result of the above, it has been possible to greatly simplify pom.xml. The maven profile "git.release" is not required. The command "mvn package" now builds all JARs.

However, the most significant change is in how pagination works. OrderBy is eliminated as a Table setting. SNDML now always orders by "sys_id" when performing REST queries. The "sysparm_offset" parameter is no longer used. Instead the code uses "sys_id>lastvalue^ORDERBYsys_id" to read through the table. This change has improved reliability and fixed issues related to large tables containing ACLs which restrict visibility of some records.

3.0.3

Minor bug fix release for:
#4
#6
#9

sqltemplates.xml change to use data type text instead of varchar for large PostgreSQL fields.

3.0.2

Fix for #3

3.0.1 - Initial Post-Beta Release

This is the initial post-beta release of SNDML3.

Changes since the beta 9:

• Default OrderBy is "sys_created_on,sys_id" for improved reliability when loading large tables.
• Action: Sync fixes.
• Changes to Logging.

Please open an issue for any bugs or comments.

3.0.1 Beta 9

Beta 9 is the first release to support action: sync.

3.0.1 Beta 8

In this release

• Fixed an issue where Partition loader aborted if zero records were processed.
• Created now defaults to [void,start]. In other words, the loader will not process records inserted after it starts running (unless you override the created setting). Also it will not process records if sys_created_on is empty. To get these records specifycreated: [void,void], filter: sys_created_onISEMPTY.
• Added orderby option
  • orderby: +sys_id - load will be ordered by sys_id, ascending
  • orderby: -sys_created_on - load will be ordered by sys_created_on, descending
  • there is not currently a default orderby field
• Added minrows and maxrows options.
  • minrows: 1 - abort if 0 rows are loaded
  • maxrows: 10000 - abort after loading 10000 rows
• Added sqlbefore and sqlafter options.
• Support for PostgreSQL.