Creating Databases
Database File Locations
A database consists of several “locations” and within each location are files:
Data location
Database 'default' location contains the following files:
• Files created when a valid user creates an Actian X table, index, or system catalog. These files have an extension starting with "t" followed by a number (for example, aaaabdef.t00), where the first extent is named t00, and succeeding extensions are named t01, t02, and so on.
Data tables contain the user data and may be referred to as data files. For details, see the chapter “Managing Tables and Views.”
System catalog tables contain metadata (for example, descriptions of tables, columns, and views) about the database and are owned by the user $ingres. For a complete description of the system catalogs, see the appendix “System Catalogs.”
• The configuration file (aaaaaaaa.cnf).
This is the most important file in the database location. If it is lost or becomes corrupt, connecting to the database is impossible. This file holds a variety of information (for example, date of last checkpoint), which can be viewed using the utility infodb.
• Files created during a modify of a table or index.
These files have an extension starting with "m" (for example, a0aadnoq.m00). These are transient files that exist only during a modify.
Database 'vectorwise' location contains the following files:
• vectorwise.conf – A default configuration file used for databases that contain X100 tables unless a database specific file is defined.
• CBM folder, which contains:
– authpass – Authorize key access file for security
– db_key – UUID for the database
– default – Folder under which all table files are created
– lock – Access file
– node_local – Distributed use only (Vector in Hadoop)
– profiles – Default folder used for the output of global profiling
• wal folder, which contains:
– location_map – Map of the system locations and X100 locations
– main.wal – Write-Ahead Log used to record all activity in the X100 database carried out by the user and triggered by configurable parameters. This can include PDT data.
– pdts – Position Delta Trees containing updates, inserts, and deletes to the X100 tables. This excludes BULK and SET-WISE operations.
When tables are partitioned, the names of the files containing the data include partition divisions.
• Files created when a valid user creates an X100 table. These files have a name format that is descriptive, containing:
– Schema name with prefix '_' (underscore)
– Table name with prefix/separator 'S' (capital S)
– Column name with a double '_' prefix ('__')
– Unique hex ID with leading zero representation
_ingresSnoparrt__first_name_0000000000000088
_ingresSnoparrt__second_name_0000000000000089
When tables are partitioned the names of the files containing the data include partition divisions. The table name is postfixed with '@' and a decimal partition number starting with zero. Example:
_actianSfivepart@0__first_name_000000000000007e
_actianSfivepart@0__second_name_000000000000007f
_iactianSfivepart@4__first_name_0000000000000086
_actianSfivepart@4__second_name_0000000000000087
Checkpoint location
Contains a snapshot of your database (entire or partial), which is created when you take a checkpoint (ckpdb) of the database.
Journal location
Contains a record of the changes made to the database’s tables and/or system catalogs. Changes to tables are journaled only if journaling is switched on for the database and the table.
Dump
Contains a record of the changes to the database table data that occurred during an online checkpoint. Dump files are used during the recovery (rollforwarddb) of a database that was checkpointed online.
For additional information about checkpoint, journal, and dump files, see the chapter “Performing Backup and Recovery.”
Work location
Contains files used for system work, such as sorting and hash joins within a piece of SQL.
Last modified date: 08/14/2024