On Tue, Aug 16, 2011 at 11:47:55PM +0530, Kashyap Chamarthy wrote:
On 08/16/2011 11:35 PM, Richard W.M. Jones wrote:
>On Tue, Aug 16, 2011 at 11:21:44PM +0530, Kashyap Chamarthy wrote:
>>I see. From a quick look at the spec file, I do notice post, preun,
>>postun operations where it deals w/ seliux policy modules, and other
>>system commands etc. Can you please elaborate a little on what kind
>>of workarounds are we talking here?
>
>You have to imagine that the appliance will boot with just the files
>specified directly from the RPMs (ie. the %files sections). Any other
>adjustments done by %pre or %post scripts will not be done.
Just a question, is there a technical/design reason why febootstrap
doesn't do this? Or like you mentioned at the end, it is purely for
more control for libguestfs ?
It's a technical reason. We don't install the RPMs in the regular way
(ie. using rpm), but instead we unpack them. We *used* to install the
RPMs in the regular way into a chroot, but it was a gigantic pain in
the rear the whole time, and we were glad to get away from doing that.
In any case, installing the RPMs like that didn't really solve the
problem, it just added a lot more complication for what is in practice
only a minor issue.
(The way forward in some distant future febootstrap 4 is to install
the RPMs in a virtual machine using kickstart or the equivalent
mechanism on other distros ...)
Meanwhile, my run of 'febootstrap' looks like still(2 hrs or
so
now) it's hung. Would you suggest me to wait? or do a ctl+c ; cleaup
all the febootstrap temp. directories; and retry?
Try the simpler example from the blog to make sure it's generally
working.
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://et.redhat.com/~rjones/virt-df/