Comment # 8 on bug 969528 from Patrick Schaaf

(In reply to Stefan Schubert from comment #7)
> Ok, thanks, but then it is not a YaST or zypper problem because the rpm
> installation has been started and has not returned back.

Sure it is not a zypper problem. We categorized this bug as component
"autoyast" for a reason :)

> So from my point of
> view it is package problem. As it is from 
> http://dl.google.com/linux/chrome/rpm/stable/x86_64 we have no influence
> here.

It is an autoyast-environment-during-install problem, as much as a package
problem.

As I wrote above, that chrome package contains a post install script that
insists on running some kind of "at" job, and to that end, nowadays contains a
call to start atd (the "systemctl start --full atd.service" we saw in the
process list) - and THAT is what is hanging, presumably because atd depends on
syslog.service, but syslog.service is not running in the autoyast environment
at that point in time.

Of course that can be classified a package problem. It did (and does) work for
our openSUSE 13.1 autoinstalls, though, so it is also a regression in the
autoyast-environment-during-install.

> Can you install the package via zypper manually after finishing the
> installation ?

Yes.

> Hm, I am still wondering why the zypper lock has not hit here....

The autoyast <init-script>s do not run with zypper locked.

We used scripts in that section already for 11.4, and more extensively for
13.1, to install the repos _we_ want, automatically run an update round, and
install packages we want additionally. Always worked fine.