On UNIX systems, disk space must be physically written when a journal file is extended. When a journal file is filled, a new one is created. It is undesirable for performance to be affected by file allocation that occurs at unplanned intervals.

You can use the alter database space preallocation features to manage when the allocation takes place, allowing control over when the allocation time delay occurs. A significant amount of journal file I/O can occur when the first journal file is created, with the archiver being unavailable during this time. This can be observed as an online checkpoint taking a long time to complete, or the archiver performing a large amount of work when the first journal write after an offline checkpoint takes place.