Kbase 19227: How to Increase Dump/Load Performance Using Binary Dump/Load?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
7/30/2009 |
|
Status: Verified
GOAL:
How to Increase Dump/Load Performance Using Binary Dump/Load?
GOAL:
How to effectively use binary dump and load?
GOAL:
How to increase performance of a binary dump and load process?
GOAL:
Can binay dump load be executed in multi-user mode?
GOAL:
Can a binary dump/load be executed online
FACT(s) (Environment):
All Supported Operating Systems
Progress/OpenEdge Versions
FIX:
The binary dump/load can be useful in certain situations such as:
- Large tables:
Binary dump/load automatically creates a new extent (.bd2, .bd3, etc) and switches to it if the dump file reaches the OS file size limit. - On-line in Multi user mode:
The database can be loaded in multi user mode. - A fast way to dump and load a database:
The binary dump/load can be faster than bulkload. You must use the command line utilities to perform the dump and load process. As with bulkload, the binary dump and load is an option of the PROUTIL command.
The syntax for dump is:
proutil <<db-name>> -C dump <<table-name>> <<path-to-bd-file>> This command generates a binary dump file (.bd) in the directory specified by the path-to-bd-file parameter.
The syntax for load is:
Proutil <<db-name>> -C load <<bd-file>> You must execute the dump and load for each table on the database. There is no option to dump and load the entire database with a single command. A good idea is use a 4GL command to read the metaschema table and automatically generate the dump and load script for all tables.
The load process generates I/O activity on the before-image file (.bi). So to avoid I/O problems caused by high number of checkpoints, increase the cluster and block size of before image file (.bi) using the command:
proutil <<db-name>> -C truncate bi -bi <<cluster-size>> -biblocksize <<bi-block-size>> Another good parameter used to make the process faster is no-integrity option (-i) on load process.
After you finish the load process, you must rebuild all indexes because binary load does not create the index structures of database.
Helpful hints:
-RO parameter for dump
-i for load
-bi and -biblocksize
Multi user mode: More than one table loaded at a time.
Database tuning parameters: Use on startup of database to increase performance on load processes (-B, -spin, -hash, -bibuf, etc.)
The build indexes option can also be run online.