• Article Type: General
• Product: Aleph
• Product Version: 18.01

Description:
How often should UTIL-X jobs be run?
Do they delete everything or can you limit them to just older files?

Resolution:
The UTIL-X-8, which runs against the VIR01 Znn tables, needs to be run at least weekly; at larger sites, nightly. It deletes everything.
How often you run the others depends on how much space you have. You can do "df -k yyyy" where "yyyy" is the disk on which the ./u50_5, $LOGDIR, and $TMPDIR directories reside.
To see the former do "echo xxx01_dev" where "xxx01" is your library. The first directory in the path is the one for which you do the "df -k yyyy".
To see the $LOGDIR, path, do "echo $LOGDIR".
To see the $TMPDIR path, do "echo $TMPDIR".
When you run p_manage_01 or other indexing jobs you will definitely need to clear these out to make work space.

Aside from that, the main consideration is how long it takes to display the list of files. You may want to create your own cron job for doing this. UTIL-X-3, UTIL-X-4, UTIL-X-5, and UTIL-X-9 have a "to-date" parameter which limits the deletion to the past xx days; UTIL-X-1 and UTIL-X-6 retain the current day; UTIL-X-2, UTIL-X-7, and UTIL-X-8 delete everything.

You should probably not run UTIL-X-3 at all: Not much is written to the libraries' $data_files directories and you can individually clean up what is written there.

In diagnosing problems with indexing jobs, Ex Libris staff may sometimes want to see the last several runs of the job. Thus, you should try to save the last logs of p_manage_nn index jobs (in $alephe_scratch) --even if they are a year old. This means copying them somewhere else or not running UTIL-X-2.
Also, it can sometimes be useful to see older run_ue_01/run_ue_08 logs (in the libraries' $data_scratch) for comparison with current.
Likewise the server logs in $LOGDIR, but these can be quite large and it may not be practical to save more than a couple days.

• Article last edited: 10/8/2013