Home Page

DBD2 Concepts

Conceptual Operational Overview
Configuration Elements

DBD2 Sample Files

Glossary of Terms

External Links

DBD2 Project Site
Syslogd2 Project Site
SLP2 Project Site

Other References

RFC 3164 (The BSD Syslog Protocol)
RFC 3339 (Internet Time Format)
RFC 5424 (Syslog Version 1)

DBD2 Operational Overview

Top of page]

Before attempting to configure DBD2 or to go into details, understanding how DBD2 works and what it does is appropriate:

When the DBD2 server receives an incoming data string from the dbcl client (note the input string is a comma-separated list of name=value pairs):

template=y1z2, user=tom. jobid=j123, time=Mar 3 2023, ...

it first parses the list of name=value pairs. Most of the name + value pairs are stored for later reference, but one special pair with the name of 'template' is used immediately. In the example string, provided, the 'y1z2' value is read to mean that two database updates will be created from this one input:
(1) Database 'y' will be updated by processing its template-script number 1.
(2) Database 'z' will be updated by processing its template-script number 2.

For each template-script that is to be processed, a starting (empty) output-string is prepared with two pointers:
(1) An input pointer that points to the beginning of the database-template to be processed.
(2) An output pointer that points to the start of the (empty) output-string.
[Top of loop]

• If the input pointer has reached end-of-string, the loop terminates and the resulting output-string is sent to the location specified by the database-identifier.
• If the input pointer is pointing to 'boilerplate text' (not a variable placeholder), that text string is copied to the output-string verbatim and both pointers are advanced to the start of the next variable placeholder.
• If the input pointer points to a 'variable-placeholder' (an sql-index variable):
  1. The value of the sql-index is used as an index into the global list of sql-variables (configured with the 'var' keyword) to get a list of names:
     var = 100 = user, username, name
     var 200 job, jobname
     ...
  2. The list of names is compared to all the names in the received data-string.
     -->If a match is found:
     • If the The associated value (from the name=value pair) is copied to the output-string and both input and output pointers are incremented appropriately.
       
     -->If a match is NOT found, nothing is copied to the output (the value defaults to a 'null-string', but the input pointer is incremented past the sql-variable.
• Processing returns to top-of-loop.

Configuration Elements

Top of page]

The overall structure of the configuration file is based on that of MariaDB with sections deliniated by section-headers in square brackets ('[ ]').

Many of the most important keywords in dbd2 actually define data mappings (a database term) that define relationships (or connections) between different variables.

Dbd2 configuration-file sections are of two types. There is one [global] section that uses the keyword 'global', and one or more output-sections with user-defined names.

In no particular order, the major components of an output-section are:

• A database-id or 'DestId': This single, non-case-sensitive alphabetic letter identifies this output-section. When combined with a template-index it creates a template-id. Multiple template-id(s) can be concatenated, producing a routing-template (or routing-string).
• One or more Database-Templates, each mapped to a template-index by the keyword 'template'.
• A connection-specification defining how to contact the database or output-file.
• Database name, type and log-on credentials if applicable.
• Threadpool settings for this output-destination

In no particular order, the primary components of the [global] section are:

• Input connection-specifications.
• Global and per-threadpool threadpool-parameters.
• Global variables and boolean values for general configuration.
• 2-way 'var' keyword mappings between sql-index<-placeholders in databaese-templates and the 'name' component of incoming name=value pairs.
• 4-way '$' keyword mappings between base-variable-types, sources of base-variables, base-variable-options, and identification-names
• 2-way 'timestring' keyword mappings between a time-string-index-number and a time-string-format
• 2-way 'syslogkey' keyword mappings between a syslog-selector-string and a template-routing-string