Red Hat Bugzilla – Bug 58886
kernel install failed using software raid
Last modified: 2015-01-07 18:54:39 EST
From Bugzilla Helper:
User-Agent: Mozilla/4.78 [en] (X11; U; Linux 2.4.9-13 i686)
Description of problem:
ran graphical up2date to install kernel-2.4.9-21 and friends. I have software
raid to mirror all my partitions. the drives are on a pci ide card (since the
via ide on the motherboard sucks). the installer bombed trying to find
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. well, i'll only try it once so it is kind of every time.
Actual Results: % up2date
Traceback (innermost last):
File "/usr/share/rhn/up2date_client/gui.py", line 786, in doInstallation
File "/usr/share/rhn/up2date_client/up2date.py", line 2148, in
bootloader = checkbootloader.whichBootLoader()
File "/usr/share/rhn/up2date_client/checkbootloader.py", line 45, in
fd = os.open(bootDev, os.O_RDONLY)
OSError: [Errno 6] No such device or address: '/dev/hda'
Expected Results: blissful silence, and everything works.
the progress bars for both the package and the total transfer both show 100%.
The package I believe it was trying to install when it bombed was
/var/spool/up2date/kernel-source-2.4.9-21.i386.rpm. the only active button on
the user interface is "cancel". and of course I don't know at this point if the
install was successful or not. (but I suspect it was.)
I see a fundamental flaw in the logic.
/usr/share/rhn/up2date_client/checkbootloader.py looks to see if
/boot/grub/grub.conf exists, and if it does then the code assumes I am using
grub. I do not use grub, when I installed I asked for lilo to be used, and
apparently the installer dropped the grub stuff in anyway. (since I asked for
"everything" to be installed.)
Created attachment 43617 [details]
Created attachment 43618 [details]
Created attachment 43619 [details]
I keep telling you guys that grub is evil. :)
and I just noticed that up2date updated the grub.conf file and not the lilo.conf
file. is that worth a separate bugzilla?
This should be fixed in the next version of the client. (checkbootloader.py
Verified by creating an /etc/lilo.conf, running /sbin/lilo, and doing an
last week I ran up2date in graphical mode to install the 2.4.18-17.7.x kernel.
several good things. /etc/lilo.conf was updated correctly. and the system
handled the software raid mirroring correctly. looks like this one can be closed.