Re: [rdo-dev] [tripleo] missing centos-8 rpms for kolla builds

On Wed, 29 Jan 2020 at 11:31, Alfredo Moralejo Alonso <amoralej(a)redhat.com&gt; wrote: > > > > On Tue, Jan 28, 2020 at 5:53 PM Mark Goddard <mark(a)stackhpc.com&gt; wrote: >> >> On Tue, 28 Jan 2020 at 15:18, Mark Goddard <mark(a)stackhpc.com&gt; wrote: >> > >> > On Mon, 27 Jan 2020 at 09:18, Radosław Piliszek >> > <radoslaw.piliszek(a)gmail.com&gt; wrote: >> > > >> > > I know it was for masakari. >> > > Gaëtan had to grab crmsh from opensuse: >> > > http://download.opensuse.org/repositories/network:/ha-clustering:/Stable/... >> > > >> > > -yoctozepto >> > >> > Thanks Wes for getting this discussion going. I've been looking at >> > CentOS 8 today and trying to assess where we are. I created an >> > Etherpad to track status: >> > https://etherpad.openstack.org/p/kolla-centos8 >> > > uwsgi and etcd are now available in rdo dependencies repo. Let me know if you find some issue with it. I found them, thanks. > >> >> We are seeing an odd DNF error sometimes. DNF exits 141 with no error >> code when installing packages. It often happens on the rabbitmq and >> grafana images. There is a prompt about importing GPG keys prior to >> the error. >> >> Example: https://4eff4bb69c321960be39-770d619687de1bce0976465c40e4e9ca.ssl.cf2.rac... >> >> Related bug report? https://github.com/containers/libpod/issues/4431 >> >> Anyone familiar with it? >> > > Didn't know about this issue. > > BTW, there is rabbitmq-server in RDO dependencies repo if you are interested in using it from there instead of rabbit repo. It seems to be due to the use of a GPG check on the repo (as opposed to packages). DNF doesn't use keys imported via rpm --import for this (I'm not sure what it uses), and prompts to add the key. This breaks without a terminal. More explanation here: https://review.opendev.org/#/c/704782.