Hardware-based backups including OS backups, disk mirroring, “snapshots,” and third party tools are not the best way to back up the database as all of these require the database to be in a quiescent state. In order to get a snapshot of the database files, you have to shut down the DB server or use proquiet. Careful with proquiet: you need to wait until the “Quiet point has been enabled” message appears in the db.lg before proceeding.
These utilities also don’t discriminate between database files and any other files in storage. If you’re trying to backup only the database, you have to know what pieces to back up and be sure the list of new/changed/ deleted files and directories is up to date. There are reasons to use these tools, but typically they involve very large (terabytes) databases.
This is the preferred method endorsed by experienced Progress DBAs everywhere.
Designed specifically to backup OpenEdge databases, it knows where all of the parts of the database are, even the parts in unusual disk locations. This means even if the database is down and OS-based utilities would work, probkup is still a more thorough, focused backup tool.
Once you specify a database, it automatically grabs the right files and includes an option to “skip” empty space in the database. In later versions of OpenEdge (11.6 – watch out for a bug in earlier implementations of this feature), only the active clusters of the BI file are backed up, again potentially reducing the size of the backup file. Probkup can also be used to change the “physical” database structure (extent size & location). Most importantly of all, probkup is the only supported method for backing up a database while it is online and in use.
This is an easy one. Always use probkup to back up all of your databases.
Example:
$ probkup online <dbname> <path><dbname>.pbk -com
In addition to backing up the database, make sure you backup the following list of directories and files.
We have a whole series of posts for new DBAs. What topics would you like to hear about first?