[Pkg-net-snmp-devel] Bug#606784: snmpd: package fails to upgrade properly from lenny

Julien Cristau jcristau at debian.org
Tue Dec 28 21:30:41 UTC 2010

On Sat, Dec 18, 2010 at 22:50:24 +0000, Jonathan Wiltshire wrote:

> (dropping the submitter, there's a deeper problem here)
> On Sat, Dec 18, 2010 at 11:01:22PM +0100, Julien Cristau wrote:
> > the postinst does:
> > 
> > if [ "x$1" = xconfigure ]; then
> >         if [ ! `getent group snmp >/dev/null` ]; then
> >                  deluser --quiet --system snmp
> >         fi
> >         adduser --quiet --system --group --no-create-home --home /var/lib/snmp snmp
> >         chown -R snmp:snmp /var/lib/snmp
> > fi
> > 
> > Which seems pretty weird.  What is this trying to do?
> Mmm, I hadn't dug this far. As it's not my package I've no idea what the
> intention was, but the call to adduser first appears in the Lenny package
> without a --group option, so the system user gets added to nogroup. This
> looks like an attempt to give the snmp user an snmp group if that's the
> case. Indeed, r286 and r293 in SVN, and bug #520724 seem to back this up.
> Being a bit green at this sort of thing, is this neccessary to ensure the
> user and group have the same ID or is it acceptable to:
>  - move the call to adduser above the test, which will fail gracefully if
>    the user already exists
>  - then test for the presence of a group snmp
>  - if there isn't, create it and add the user to it?
Not just add the user to it, but set it as the user's initial group.  So
'usermod -g snmp snmp' or so.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-net-snmp-devel/attachments/20101228/93d02fa9/attachment.pgp>

More information about the Pkg-net-snmp-devel mailing list