begin  quoting John H. Robinson, IV as of Tue, Aug 22, 2006 at 02:32:43PM -0700:
> Stewart Stremler wrote:
> > 
> >  5. Redirect all of the output to /dev/null, and forget about it. You
> >     should have a way to check that your scripts ran w/o checking your
> >     email, so the email is just clutter.  If you discover a problem,
> >     you can rerun the job and look at the output.
> 
> This works only if the problem is reproduceable at the time the script
> is run by hand. If the script s faling due to network connectvty because
> a router is getting rebooted, every day, at the same time your script is
> running, then you won't be able to reproduce the error.

True. But if it's a one-time failure, there's not much to do. And if
it's a regular failure, you undo the redirect, and check the email every
day for awhile.

/me shrugs

If it's a system that's critical for several people, that's one thing.
If it's a system that can go down for a week at a time w/o anyone 
noticing, that's something totally different. Like I said... "it
depends". :)

> I would think the better would be to log the output somewhere, rotate
> it, and use the alternate means of ensuring the program ran
> appropriately. This way, when there is a problem, you can go to the logs
> and see what happened *when* it failed.

I actually didn't see the problem with just putting up with the email
from each task.  I just figured that there was more than just two ways
to handle the problem. :)

> Now, if the script produces something useful at that time is anyone's
> guess.

There's that.

-- 
_ |\_
 \|


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to