Content Server Manual / Version 2310
Table Of ContentsThe CoreMedia Cleaners are tools for freeing up memory on the server and the underlying database. They are command line tools with no graphical user interface and are intended for use by system administrators only. Administrators are advised to run these collectors regularly (by cron jobs, for example) in order to maximize efficiency and speed of the repository. They come in two versions:
Clean Recycle Bin: A deleted resource is moved into the Recycle Bin. To remove the resource from the Recycle Bin use the Clean Recycle Bin utility. It will remove a resource A permanently from the Recycle Bin if and only if there is no resource B outside the Recycle Bin that is linked to A. This ensures that no dead links exist in the system (or on a website that is based on the content). On high-performance systems this collector should be run once a week.
Clean Versions: The Clean Versions destroys unnecessary intermediate versions of content items. Flexible parametrization allows for arbitrary version selection. The cleaner will not alter the servers runlevel. By providing some folders which are then processed recursively it is possible to divide the cleaning workload across multiple collection jobs.
Both types of cleaners are closely tied - although the actual processing is quite different. Imagine two types
of content items: Navigation and Article. Navigation content items point to
Articles with some LinkList Property. Articles change daily and are linked to a single navigation
content item. As long as some version of the navigation content item still points to an article, this article cannot
be collected out of the Recycle Bin. The solution is to run the
Clean Versions in order to delete the intermediate versions of the
navigation content item. This will also remove the old links to articles in the
Recycle Bin - and thus allow the content item collector to finally
remove these content items from the Recycle Bin.
This is the reason why Clean Recycle Bin and Clean Versions should be run regularly. Both types of cleaners run in a distributed system and are processing and IO intensive. Actual running time depends heavily on network latencies, database performance and number of resources processed.
