slocate doesn't

inode0 inode0 at gmail.com
Sat Apr 15 16:18:04 UTC 2006


On 4/15/06, Michael Johnson <mjohnson at coderyte.com> wrote:
> On Apr 15, 2006, at 9:10 AM, Dag Wieers wrote:
>
> > Even on production servers you don't want to have it running. We
> > seldom
> > touch those anyway once in production.
>
> Just curious as to why.  I'm a diversity sort of guy.  I don't like
> the single point of failure idea of all the servers being exactly the
> same brand, disk topology, processor, and OS.  Some things you can't
> get around, but for the most part keeping diversity helps when
> something goes seriously wrong.
>
> With that in mind, when I'm looking for my httpd.conf file, it might
> be in /etc/httpd/ or /etc/httpd/conf/ or even in /usr/local/etc/
> httpd/ among others.  I know the file I'm looking for, but on
> different systems it can be in slightly different locations.  For
> this purpose, I like to have slocate running.  It's not that much of
> a drain on the system at all.

On a production server it shouldn't be changing that often in
substantial ways. So set the server up and run updatedb to get an
accurate snapshot of the local filesystem. Then you can rerun updatedb
as needed. I usually run it after upgrades and/or significant updates
are applied and when I install new software (infrequent on a
production server).

>From day to day it just shouldn't be changing ... so running it once
in a while if you use it should be enough. Some people want to run it
every day and I wouldn't argue with them about doing it ... I just
don't see any need to rebuild the database anywhere near that often.

On my desktop and development machines where I'm experimenting and
have a volatile filesystem I do run updatedb more frequently, on some
daily and others weekly.

John




More information about the redhat-list mailing list