Red Hat Bugzilla – Bug 88983
post 9 install -- up2date crashes system
Last modified: 2007-04-18 12:53:06 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; T312461;
Q312461; .NET CLR 1.0.3705)
Description of problem:
Red Hat 8, ext3, upto date per 'up2date' system updated to 9 via CD burned from
Install -- OK.
Red Hat Network (up2date), post install listed number of packages
needed to be updated. Made several attempts -- all 'successful' but
up2date says still needed to run.
Manual check should vi rpm -q show version differences.
Ran 'rm -f /var/spool/up2date/*'. This was a suggested solution
to a previous reported bug for a failed up2date (php-manual).
Attempted to run up2date. This attempt hung system. Cold reboot
results in unbootabel system:
Booting 'Red Hat Linux (2.4.20-8)'
Filesystem type is ext2fs, partition type 0x83
Kernal /vmlinuz-2.4.20-8 ro root=LABEL=/
[Linux-bzImage, setup=0x1400, size=0x10a4f4]
[Linux-initrd @ 0x7fb5000, 0x3ae76 bytes]
Uncompressing Linux ...
-- System halted
I am VERY, VERY, VERY upset.
Version-Release number of selected component (if applicable):
unable to provide -- most recent ?
Steps to Reproduce:
Actual Results: System unusable.
Expected Results: Clean update.
See registered system 'pglm' for system details.
I am able to boot from floppy.
Doing a ls -l of /boot I find two files "vmlinux-2.4.20-8".
Two files same name, same directory, same attributes, different
One is 3036023 and the second is 1096436 ??
How is the file manager allowing duplicate names?? Which is correct??
Also on another issue -- lpd fails start on boot "Cannot bind to lpd port'515'".
In searching the net I find others with the same issue -- what gives?
The file manager is not allowing two files with the same name. This appears to
be file system corruption! You need to do an exhaustive file system check (via
e2fsck if your file systems are of the default ext3) on your /boot partition.
I don't think this bug has anything to do with up2date, however. Until e2fsck
gives the thumbs up on your system's computer - you may not be able to figure
out the problem.
I used e2fsck -n on all mounted partions.
All reported "clean".
"clean" just means e2fsck checked to see if the clean bit was enabled - it was.
"clean" status is an assumption that the file system has been checked, fixed,
and validated. You must "-f" (force) e2fsck to ignore the clean status.
1) my mistake eyes aren't what they should be on file dup --
one is vmlinux..., other vmlinuz....
2) e2fsck -f ... no errors.
3) reboot from floppy -- ok
4) reboot from / (/dev/hda5) -- looks ok
5) Now -- why did up2date not update -- downloads rpm's but not apply
(the orginal question)?
Trying to update --- again.
Red Hat Network says updates available, and is trying to update now.
Icon is blue check mark -- Up to date indicator?
Running Red Hat Network (up2date).
Found packages, check conflicts, downloaded --
it has now been installing for over 5 min. --
hour glass spinning, progress bars not showing
up2ate just silently died.
Icon red. Clicking on it -- notification tool
indicates updates not applied.
Yes there is a problem!!
Did a e2fsck -f /dev/... for all (hda1...).
I have run from the commandline :
1) up2date -p
2) up2date -v -i -k /var/spool/up2date/*
3)rpm -i /var/spool/up2date/samba-swat-2.2.7a-8.9.0.i386.rpm
file /usr/sbin/swat from install of samba-swat... conflicts with file
from package samba-swat-2.2.7a-7.9.0
file /usr/share/man/man8/swat.8.gz from ... conflicts with file ...
Same result for balsa.
Running rpm -v -U /var/spool/up2date/balsa-2.0.6-2.i38b.rpm
did get the package to install.
Ergo -> up2date is not working??!??
Ver 8 used SWAT and had it as a menu item.
Ver 9 does not have SWAT as a menu item, /spool/up2date does not
list redhat-config-samba in menu or /spool/up2date -- net result
is the samba configurator is "lost" during upgrade -- what else
is lost? and how to "recover" ?
Orginally the ver 8 system was install with "everything".
I also do not want to install all of the kde and spdf non-English language