[509] in testers
Re: named database reloading
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Fri Dec 15 19:23:44 1989
From: probe@ATHENA.MIT.EDU
Date: Fri, 15 Dec 89 19:23:18 -0500
To: jik@PIT-MANAGER.MIT.EDU
Cc: testers@ATHENA.MIT.EDU
In-Reply-To: Jonathan I. Kamens's message of Thu, 14 Dec 89 23:10:51 -0500,
Reply-To: Richard Basch <probe@ATHENA.MIT.EDU>
Date: Thu, 14 Dec 89 23:10:51 -0500
From: Jonathan I. Kamens <jik@PIT-MANAGER.MIT.EDU>
Sender: jik@PIT-MANAGER.MIT.EDU
Richard writes:
>Actually, "kill -HUP" has never really worked properly. The problem
>this time is that your cache still has various entries listed and the
>TTL has not yet expired. This is why it concatenated the information.
>
>The man-page never claims that it flushes the database when you send it
>a SIGHUP, though I admit, it is ambiguous given a familiarity with the
>previous behavior.
Yes, it is ambiguous. Therefore, it should be documented in the man
page. The man page should state exactly what sending a SIGHUP to the
running named process will do.
I don't know if this was the conclusion you reached after my last
message, because your last paragraph above leaves me wondering whether
you agree that the man page should be changed or not. I think it
should. So, I'm asking, just so that I am clear about what you are
saying, are you considering this a bug in the man page that you plan
on eventually fixing, or are you ignoring it?
jik
Since it can be argued where the bug is, we are simply going to say that
the manpage should be fixed... However, we are avoiding changes to the
packs at this late date, since we are supposedly going to staff-test as
of Monday.
If by the building of the next release, a new version of BIND has not
come out that addresses this problem (either in the program or in the
man page), then we will document it in the man page.
-Richard
