(This was originally reported as bug #702402 on bugzilla.novell.com)
When installing a group of packages the sfcb service fails to shut down causing
some of the packages in the group to fail to install. We run into this when
building a SLES 11 SP1 appliance image.
So the problem here is that sblim-cmpi-network's post install script calls the
shell script provider-register.sh, which among other things, does this:
sfcb_rebuild()
{
if ps -C sfcbd > /dev/null 2>&1
...
/etc/init.d/sfcb stop // killproc sfcb
while ps -C sfcbd > /dev/null 2>&1
# wait 10sec and print the timeout message the customer is seeing
the real bug here is that we shouldn't be killing a process that's
running outside the chroot.
Proposed patch: Use "checkproc" instead of "ps -C"
Actually, I'd like to move provider-register.sh out to a separate package and have all providers require it (instead of bringing their own copy)