Re: Last chance for d-i changes in stretch
On 05/26/2017 09:30 PM, Cyril Brulebois wrote:
>> I've looked at that for a bit, and found out that this is my own
>> fault: one of the uploads of open-iscsi I did before the freeze
>> changed the logic on how the initiatorname was generated within the
>> installer, (due to feedback from Ubuntu people IIRC) ensuring that
>> /etc/iscsi always contains files, and the check in the finish-install
>> script now thinks that iSCSI is used in _all_ installations. (It
>> checks for /etc/iscsi/* .) For this reason it regenerates the
>> initramfs on all installations, which takes a couple of seconds.
>> The effect here is totally harmless (it just unnecessarily calls
>> update-initramfs -k all -u), but it might be annoying.
>> KiBi: I suspect you don't want to change this so close to the
>> release, but in case you do, I'd be happy to upload a targeted fix
>> for that.
> I think it would be best to postpone considering such a fix for the
> first point release, instead of aiming for r0. Let's look at this once
> r0 is out, so as to avoid generating noise for the release team? I've
> added an item on my d-i task list so that I don't forget about it.
Sure, perfectly fine with me. If I don't open a p-u bug after
the release of Stretch myself, feel free to ping me.
(Btw. I also just noticed from reading the code that the
additional time is not the only side-effect: it will clutter
every new installation with a small file in /etc/iscsi. The
file is harmless, and won't cause any problems, but I wanted
to mention it so you know about it.)