Red Hat Bugzilla – Bug 609374
pv guest install failed caused by anaconda bug
Last modified: 2012-10-04 02:08:08 EDT
Description of problem:
PV guest won't be installed when doing virt test, see installation log here
This may be caused by bug 517051, "xconfig --startxonboot" in kickstart file may lead to headless system failed in installation.
A possible workaround can be found at bug 517051 comment 4
In RHTS kickstart has no "xconfig --startxonboot"
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Submit a job installing headless system as pv guest
Or just clone job 2513
fail to install pv guest
pv guest can be installed as in RHTS
Here is a failed job not in a virtual environment
More failed jobs
This has been discussed on today's Beaker meeting. "SkipX" fix is easy to do, we just need to make sure that we're not introducing regressions elsewhere (for other teams).
The next step is to have the testing packages on beaker-stage, where people can give it some hands on to see if everything works as expected.
Gurhan, any thoughts?
Guan Eryu, bug 517051 is ON_QA.
Can you give it another shot (with the updated packages) if this is still an issue for you?
(In reply to comment #1)
> This has been discussed on today's Beaker meeting. "SkipX" fix is easy to do,
> we just need to make sure that we're not introducing regressions elsewhere (for
> other teams).
> The next step is to have the testing packages on beaker-stage, where people can
> give it some hands on to see if everything works as expected.
> Gurhan, any thoughts?
Yes indeed. skipx is the most appropriate fix here. I think, in fact I am pretty sure, we have used it in many (all?) guest kickstarts in RHTS, not sure what in cobbler have put in xconfig line. It should be ok to put in skipx in all guests' kickstarts.
(In reply to comment #2)
> Guan Eryu, bug 517051 is ON_QA.
> Can you give it another shot (with the updated packages) if this is still an
> issue for you?
I think I couldn't test it with the updated anaconda, as the virtual workflow has to install 5.5 GA version first then install new kernel in guest(I want to test new kernels in guest).
this bug got the target milestone set to 0.5.59. Did it just slip through the cracks when bugs got switched to modified, on_qa, closed - or the patch isn't in at all?
This is one of the bugs that is tracked as a SignOff Blocker (and suggested as a BeakerBlocker).
The ticket mentioned in comment #6 has been resolved. Resolving this bug too as there seems to be no negative feedback.
Please reopen if the issue is not resolved according to your needs.
(In reply to comment #8)
> The ticket mentioned in comment #6 has been resolved. Resolving this bug too as
> there seems to be no negative feedback.
> Please reopen if the issue is not resolved according to your needs.
This week, many of my virt testing jobs failed due to pv or hvm guest installation failure. And their symptoms are different from Comment 0.
Could you have a look at above jobs and logs in them?
Assuming this is now fixed.