Hi,
On 7/9/20 1:06 PM, Neal Gompa wrote:
On Thu, Jul 9, 2020 at 7:02 AM Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
Hi,
On 7/8/20 12:01 PM, Vendula Poncova wrote:
On Fri, Jul 3, 2020 at 2:07 PM Hans de Goede <hdegoede@xxxxxxxxxx <mailto:hdegoede@xxxxxxxxxx>> wrote:
Hi,
On 6/30/20 6:59 PM, Brian C. Lane wrote:
> On Tue, Jun 30, 2020 at 01:53:19PM +0200, Hans de Goede wrote:
>> So 2 questions:
>>
>> 1. What do you think of my proposal to disable these
>> services (when not needed) in the livecd post-install
>> config phase? Would you be willing to accept a
>> merge-req for this?
>>
>> 2. Its been a long time since I last touched the
>> anaconda code, my python is quite rusty and my
>> plate is way too full, so I was wondering if one
>> of you could implement this if we chose to go this
>> route?
>
> I think implementing this may be even easier. Anaconda runs post-script
> snippets from /usr/share/anaconda/post-scripts/*ks at the end of every
> install. The livecd could drop in a script to handle disabling these, I
> think.
>
> Current scripts are here:
>
> https://github.com/rhinstaller/anaconda/tree/master/data/post-scripts
Yes that would be an option, although getting the list of pkgs
which are deemed as being necessary for the configured storage
(and thus should not have their services disabled) might be
tricky to do from such a post-install script ?
Anyways, going by Vojtěch's reply, there are no objections to
removing device-mapper-multipath from the livecd. And it will be
quite easy to make the dmraid-activation service disable itself
if no dmraid sets are found (it already is a shell script).
So I believe it would be best to move forward with this as
I original proposed in the 2 change pages:
https://fedoraproject.org/wiki/Changes/DisableDmraidOnFirstRun
https://fedoraproject.org/wiki/Changes/RemoveDeviceMapperMultipathFromWorkstationLiveCD
This requires almost no anaconda changes as mentioned in another
part of the thread, the hard-dep on device-mapper-multipath needs
to be dropped. But with some luck that is all that is required
on the anaconda side.
The anaconda's dependency on device-mapper-multipath is caused by fcoe-utils:
anaconda -> anaconda-install-env-deps -> fcoe-utils -> device-mapper-multipath
Is it expected that the support for FCoE devices will be dropped from Live OS?
That was not part of the Change proposal, but having FCoE support in the
Workstation LiveCD makes about as much sense as having multipath support, so
I believe that dropping it is fine. I can amend the Change proposal to mention
this.
If yes, it shouldn't be a problem to change the dependency on fcoe-utils to weak.
If you can do that, then that would be great. But I believe that dnf includes
weak deps by default. I must admit that it is a long time since I've done
anything wrt livecd composes. Is there a way to explicitly exclude some weak
deps when doing composes, or are they excluded by default now a days?
I guess I should just re-learn how to do composes and do a compose to make
sure things turn out as we want. Is there a short primer on how to do a livecd
compose these days somewhere?
If you set them as excluded in the live media kickstart files, they'll
be properly excluded. I fixed this some time ago so that even weak
installs would not take effect in this scenario.
So you'd probably want to add a "-fcoe-utils" line to %packages in
fedora-live-base.ks:
https://pagure.io/fedora-kickstarts/blob/master/f/fedora-live-base.ks
Ok, this sounds good. Anaconda-team, can one of you turn the fcoe-utils
dep into a weak-dep then and let me know when a build with that change
has been done.
Once that is in place I will make and test the necessary change to the
fedora-live-base.ks file.
Regards,
Hans
_______________________________________________
Anaconda-devel-list mailing list
Anaconda-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/anaconda-devel-list
