Report information
The Basics
Id:
35956
Status:
rejected
Worked:
30 minutes
Users:
sar: 30 minutes
Priority:
Low/Low
Queue:

People
Owner:
Nobody in particular
Cc:
AdminCc:

BugTracker
Version Fixed:
(no value)
Version Found:
(no value)
Versions Affected:
(no value)
Versions Planned:
4.3.2
Priority:
P1 High
Severity:
S3 Low
CVSS Score:
(no value)
CVE ID:
(no value)
Component:
(no value)
Area:
(no value)

Attachments
0001-Make-No-subnet-declaration-for-.-error-info.patch

Dates
Created:Fri, 09 May 2014 08:42:22 -0400
Updated:Fri, 07 Jul 2017 20:18:55 -0400
Closed:Thu, 20 Nov 2014 01:07:08 -0500



This bug tracker is no longer active.

Please go to our Gitlab to submit issues (both feature requests and bug reports) for active projects maintained by Internet Systems Consortium (ISC).

Due to security and confidentiality requirements, full access is limited to the primary maintainers.

Subject: make 'No subnet declaration for ...' error -> info
Date: Fri, 09 May 2014 14:42:17 +0200
To: dhcp-bugs@isc.org
From: Jiri Popelka <jpopelka@redhat.com>
This is not really a bug, just a suggestion. dhcpd(8) says: COMMAND LINE The names of the network interfaces on which dhcpd should listen for broadcasts *may* be specified on the command line. This should be done on systems where dhcpd is unable to identify non-broadcast interfaces, but *should not be required* on other systems. If I don't specify interfaces to listen on, dhcpd listens only on interfaces, for whose it finds subnet declaration in dhcpd.conf. For each interface where's no corresponding subnet declaration, dhcpd spits out an *error* "No subnet declaration for <interface>. Ignoring requests on <interface>. If this is not what you want, ..." This should not IMHO be an *error*, but only *info*, because there's quite a big chance that this is exactly what user wants. If there's no matching subnet declaration for *any* interface dhcpd spits our fatal error "Not configured to listen on any interfaces!" which is all right. With regards, Jiri Popelka Red Hat, inc.

Message body is not shown because sender requested not to inline it.

On Fri May 09 12:42:22 2014, jpopelka@redhat.com wrote:
> This is not really a bug, just a suggestion.
>
> dhcpd(8) says:
> COMMAND LINE
> The names of the network interfaces on which dhcpd should
> listen for broadcasts *may* be specified on the command line. This
> should be done on systems where dhcpd is unable to identify
> non-broadcast interfaces, but *should not be required* on other systems.
>
> If I don't specify interfaces to listen on, dhcpd listens only on
> interfaces, for whose it finds subnet declaration in dhcpd.conf.
> For each interface where's no corresponding subnet declaration, dhcpd
> spits out an *error*
> "No subnet declaration for <interface>. Ignoring requests on
> <interface>. If this is not what you want, ..."
> This should not IMHO be an *error*, but only *info*, because there's
> quite a big chance that this is exactly what user wants.
>
> If there's no matching subnet declaration for *any* interface dhcpd
> spits our fatal error
> "Not configured to listen on any interfaces!"
> which is all right.
>
> With regards,
> Jiri Popelka
> Red Hat, inc.

I think your suggestion is reasonable, the difficulty is dealing with the historical
inertia of the system.  I'm not sure I want to change this from an error in case
people have built that assumption into their systems (nanny scripts and the like).
We shall think about this some more and figure out if there is a clean way to
handle the change.

Thanks for the suggestion,
Shawn


As you may have seen I sent a question about this issue to the users mailing list.
The current view seems to be that it isn't needed to change this, but nobody has
said changing it would be a problem.  I still haven't decided what we shall do about it.

You may wish to contribute any comments of yours to the mailing list.

regards,
Shawn

As you may have seen I asked about this on the
dhcp users mailing list.  The response was mixed
which I interpret as not sufficient to make this change.

I do think it probably would have been the right thing
to use info instead of error at the start but now that
it has been in use for years I don't think it would be
a good idea to change it.

As always thank you for the suggestion and please
keep sending them in.  We are trying to respond to
them a bit more quickly and are trying to work through
some of the backlog.

regards,
Shawn