rawhide report: 20050417 changes

Jeff Spaleta jspaleta at gmail.com
Mon Apr 18 22:11:36 UTC 2005


On 18 Apr 2005 18:54:34 -0300, Alexandre Oliva <aoliva at redhat.com> wrote:
> Uhh...  You mean, a bit more than 1 hour past boot up time, right? 

You tell me what i mean. I don't notice anacron.. i don't have systems
that I reboot on a regular basis... that don't invovle a coffeebreak. 
I'm reusing terms and phrases other people have used to define the
problem associated with slocate so far in the sporatic discussion on
the issue. I'm still desperately trying to understand exactly why
people who do regular system boots want anacron running.. at all. If
slocate as spawned by anacron causes a problem.. an hour after
bootup.. surely prelink is going to be noticable too as well as other
default scripts.

For the sake of simplicity, let's stick to laptop user experience,
since laptop users have been a primary source of first person
complaints about slocate as managed by anacron in these list
discussions.  Out of all of the scripts anacron is configured to run
periodicly, which ones are really crucial to run and thus something
laptop? Do laptop users need anacron running at all? Is running
prelink an hour after bootup on a laptop, like anacron does really
beneficial?  Is anacron a net negative for laptop users or a benefit
its default configuration?  Looking beyond default daily cronjobs...
how many packages in Extras or Core or 3rd party packagers drop
scripts in cron.daily designed primarily to run at 4am in the morning
without a thought to anacron's 1 hour delay behavior?

I think the slocate complaints so far are just symptoms of a deeper
problem with what anacron tries to do in general and how anacron's
default configuration tries to mimic vixie-cron behavior by running
ALL the same scripts as vixie-cron in a delayed manner. Anacron just
wasn't designed to really deal with laptops, but so far most of the
complaining in these discussions have been from laptop users.

-jef"which users does anacron really serve?"spaleta




More information about the fedora-test-list mailing list