how much beer do I need to get this patch applied?

Kurt Buff kurt.buff at gmail.com
Wed Jun 20 19:56:47 UTC 2007


On 6/20/07, Jo Rhett <jrhett at svcolo.com> wrote:
> On Jun 20, 2007, at 10:40 AM, Kurt Buff wrote:
>
> > On 6/20/07, Dan Rue <drue at therub.org> wrote:
> >> If it's too broken to complain, then the behavior is the same with or
> >> without this patch.
> >
> > Indeed, which is why this patch might not be such a good idea. In this
> > case, absence of evidence is indeed evidence of absence, which is
> > contrary to the general case.
>
> You appear to be completely confused about what this change does.
> All it does is TO ALLOW (not require) the OP to disable the spurious
> and empty output from successful cron jobs.
>
> If I get a message every day saying "No output", how do I know when a
> failure has occurred?  This patch changes nothing about that
> behavior.  Getting no message is equally useless in the situation
> where no output was generated *AND* the result code is positive.

Currently, if you get no message from that box, *something* is broken.

> The more likely is that the OP starts deleting the messages unread
> each day and thus never sees an actual failure report.

Failure of imagination.

> > Perhaps the OP needs a better way of dealing with the notifications
> > than simply turning them off.
>
> How do you suggest dealing with 1200-1800 messages which simply say
> "no output" each day?  The commands were successful, and the
> processes had no output.
>
> 1. In that load level I won't notice one missing, so absence of the e-
> mail is not useful.
>
> 2. In that load level I can't possibly read them all.  So actual
> reports of failure will be overlooked.
>
> 3. Actual errors *will* be reported, and *will be read* if I don't
> have to delete thousands of non-errors.

Perhaps a separate mailbox dedicated to this task, with a script
(grep?) that parses the emails in that mailbox daily looking for
expected messages, noting and deleting them, with unsent messages
noted via an email and messages with unexpected content forwarded as
well?

Kurt


More information about the freebsd-stable mailing list