Update Database Schema
Kamailio's database schema is specified in an XML format, which allows generation of creation scripts for many database types at once.
The XML files with the database table definitions are located in the source tree, inside subdirectory:
Each table is defined in the file tablename.xml. For example, the definition of table domain is in file:
Changing the database schema (adding new tables, adding/removing/updating columns) must be done creating or updating the XML files.
Regeneration of the database creation scripts is done running the next command inside the root folder of the source tree:
This command generates the SQL or control files to be used by kamdbctl tool when creating the database structure. It requires the tool xsltproc to process the XML files, tool which is also needed for generation of the readme files. The sql scripts will be placed in subfolders of 'utils/kamctl/'
Changing the structure of an existing database table requires only the above steps. Adding a new table requires as well:
- adding the table to be part of a specific group of tables. If the table is used by a module that uses other tables, all of them have to be in the same group. Tables group definitions are stored in the files kamailio-groupname.xml . It is common to use the module name as the group name.
- adding the table to kamctl and kamdbctl tools for creation
For example, the group of the tables used by module domain, domain and domain_attrs, is specified in file:
To generate the documentation of the existing kamailio-db-devel tables:
cd lib/srdb1/schema; make docbook-xml cd doc/databases/kamailio; make html firefox tables.html
If you add a new table, you must add the new files related to it to GIT repository, before committing the changes. It is recommended to split the commit in two parts, because they affect different components:
- one for the XML files (they are part of internal library srdb1)
- one for DB creation scripts (they are part of kamctl tool)
This brings the benefit of easy backporting, because the commit to xml file is unlikely to create conflicts, changing one file in a group of lines related to same column, so it can be cherry-picked alone. Commits to database creation scripts affects many files and can result in conflicts if new columns were added meanwhile to the development version to that table.
If there is a fix that has to be backported, always do it to master branch and then cherry-pick it in the stable branches. With split commits, the typical backport procedure is:
- cherry-pick only the commit to xml files
- regenerated the schema for the stable branch and commit the updates