Raising severity of rcS bugs that are confirmed to cause loops?

[Felipe Sateler]

On 5 November 2015 at 12:11, Michael Biebl <biebl at debian.org> wrote:
> Am 04.11.2015 um 19:41 schrieb Felipe Sateler:
>> Hi,
>>
>> I was thinking we should raise severity of the bugs regarding runlevel
>> S where we have confirmation that the bugs have caused dependency
>> loops. Currently the bugs are at severity important.
>>
>> So far I'm aware of only 2 of them: #796611 (ferm) and #796635 (nvi,
>> confirmed in LP#1489939, has full patch ready).
>
>> What do you think about rasing the severity to serious? I believe that
>> is appropriate severity for "known to break the boot".
>
> Sounds ok to me. But I was of the impression we had much more rcS init
> scripts which could dep loops. At least under certain configurations.

Right, in principle all services that do Required-Start: $remote_fs in
runlevel S can cause such loops. However, what I'm proposing is to
raise severity for those that have had at least one user say so.

> I remember nfs-common/rpcbind, ifupdown in combination with firewalld or
> NetworkManager, setserial and that's just from the top of my head.

Hm, obviously I skipped the elephant in the room :/ Setserial is
already fixed though.

I guess then the severity should also be raised in #748074 and #796637.

The first one is tagged patch, but the second isn't (didn't ubuntu
have units for nfs already?).

>
> But yeah, let's start with the packages we know for sure are
> problematic. Ideally we should send a patch along with it, for those
> that don't have one yet. At least for those packages which are important
> / aren't leaf packages.

Nod.


-- 

Saludos,
Felipe Sateler