Closed TheChymera closed 4 years ago
Yes, there was a reason, namely that we do not want the specialized systems to automatically sync the /usr/portage, because this gets bind-mounted from the stemgentoo (It's considered bad style to sync a portage repo multiple times in a row)
Ok, that makes a lot of sense, but one effect of that is, that the general-purpose OpenStack systems do not behave as expected when running e.g. eix-sync
or emerge --sync
. Any ideas how this can be addressed rather than requiring yet another manual configuration step when starting out with one of our images?
Probably just sed it away ('/auto-sync=no/d'
) in a scripts/openstack_image/*.sh
step?
The commit is only a partial fix, since it's restricted to the openstack_image
command. Ideally this would be done more generally so that it applies to all the image variants.
Then we probably should sed it in in scripts/update/default/xxx.sh and away in scripts/update/default/zzz.sh.
Or rather, do it just before and after issuing emerge --sync, id est: here https://github.com/IBT-FMI/gebuilder/blob/c2fffa772c8223c15ab90181342ed13a297d3fdf/gebuilder/scripts/common/sync.sh#L10
Is emaint sync -a
in this file really the only sync we want to protect against?
I think so, yes.
@Doeme , is there any reason we set auto-sync to no here?