The problem isn't really grub,
that's just more of a symptom. For some reason, when I run anaconda
with x86_64, when anaconda goes to add packages to the transaction set,
it adds both the i386 RPM, and the x86_64 RPM, even though there's no
i386 packages on the system This results in file conflicts when it installs the RPMs from the transaction set, obviously. To make the problem even weirder, the problem only occurs when I run the python script below (to determine hard-drive partition sizes). I'm using a fairly old version of anaconda, 11.1.2.87, to be exact. I'll check out a more recent version, but I hate reworking the scripts I have, as it's extremely time-consuming to go through the buildinstall script and figure out the new way of calling it to build an iso. Matt Joel Granados wrote: There was a temporary issue where grub was not present in one of the composes for rawhide. Your issue could be related to that. I would try again in a couple of days. regards. On Wed, Mar 04, 2009 at 05:28:43PM -0500, Matt Rose wrote:I'm having an unusual problem. I recently re-wrote my ks.cfg %pre script in python, and now anaconda can't install 64bit. 32bit installs fine, but 64bit fails with a file conflict. I put some debug code in an updates.img, and I got the following error message in anaconda.log 20:33:27 ERROR : [('file /sbin/grub conflicts between attempted installs of grub-0.97-13 and grub-0.97-13', (6, '/sbin/grub', 0L)), ('file /usr/bin/mbchk conflicts between attempted installs of grub-0.97-13 and grub-0.97-13', (6, '/ usr/bin/mbchk', 0L)), ('file /usr/share/doc/syslinux-3.11/sample/fd.o conflicts between attempted installs of syslinux-3.11-4 and syslinux-3.11-4', (6, '/usr/ share/doc/syslinux-3.11/sample/fd.o', 0L))] So I look further up the anaconda.log, and I see this: 20:33:16 DEBUG : Member: grub.x86_64 0-0.97-13 - u 20:33:16 DEBUG : Adding Package grub - 0.97-13.x86_64 in mode u 20:33:16 DEBUG : Member: grub.i386 0-0.97-13 - u 20:33:16 DEBUG : Adding Package grub - 0.97-13.i386 in mode u and then it occurs to me that all packages are being added twice, once for i386, and once for x86_64. This is my new %pre script, and it doesn't matter whether it's a separate script that I put on the CD image, or called in the ks.cfg with "%pre --interpreter /usr/bin/python" Judging from the amount of patches flying by, I know you guys are busy, but if somebody can think of a reason why anaconda would be adding nonexistent, conflicting i386 packages because of a %pre script, I would greatly appreciate it. import sys,os,time,shutil,parted,math,subprocess sys.path.append("/usr/lib/anaconda") import isys,iutil #if len(isys.hardDriveDict()) != 1: # sys.exit(123) #else: drive = isys.hardDriveDict().keys()[0] print drive if drive == 'cciss/c0d0': drive = 'cciss/c0d0p' print drive dev = parted.PedDevice.get("/dev/%s" % drive) sizeMB = (float(dev.heads * dev.cylinders * dev.sectors) / (1024 * 1024) * dev.sector_size) print sizeMB sizeMB_i = int(math.floor(sizeMB)) print sizeMB_i if int(math.floor(float(sizeMB * .1))) > 4000: backup_size = int(math.floor(float(sizeMB * .1))) else: backup_size = 4000 log_size = int(math.floor(float(sizeMB * .05))) print "/var/log size is %i " % log_size print "/backup size is %i " % backup_size root_size = (sizeMB_i - backup_size - log_size - 128 - iutil.swapSuggestion() [1] ) print "/ size is %i " % root_size ptext = """zerombr yes clearpart --all part /backup --fstype ext3 --size %i part swap --recommended part /boot --fstype ext3 --size 128 part / --fstype ext3 --size %i part /var/log --fstype ext3 --size %i """ % (backup_size,root_size,log_size) print ptext partinfo = open("/tmp/partinfo", 'w') partinfo.write(ptext) rc = partinfo.close Matt_______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list-- Joel Andres Granados Brno, Czech Republic, Red Hat. _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list |
_______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list