Skip to main content.
home | support | download

Back to List Archive

Re: Automatic Updating of swish-e Index

From: Bill Moseley <moseley(at)not-real.hank.org>
Date: Fri Jul 20 2001 - 15:23:08 GMT
At 02:03 AM 7/20/2001 -0700, Una Cullen wrote:
>Hi,
>
>Could anyone tell me if it is possible to get the swish-e index to
>update automatically, say every hour?
>I would like to be able to do this as my content is updating frequently
>and I do not want to run the swish-e command all the time.

In addition to cron as SRE explained, here's a few comments.

- If indexing doesn't take too long, you can just re-run indexing.  If you
index to a temporary file, and then rename that file after indexing has
completed you won't interrupt searches that happen while indexing is in
progress (that is, while the index file is being written).  The dev version
now writes to a temporary file and then renames the index files automatically.

- You can use the -N switch in the development version to only index files
that are newer than the file specified with the -N switch. 

- I have also used a system where an application was creating new files,
and  when a new file was created it also wrote a symlink to that file in a
parallel directory structure.  Cron ran every five minutes and indexed the
files in the parallel symlink directory and searches used -f full.index
symlink.index.  Then once a day cron ran and if there were more than X
number of symlinks, then full indexing ran and the symlinks were removed.



Bill Moseley
mailto:moseley@hank.org
Received on Fri Jul 20 15:23:26 2001