From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0)
Description of problem:
During RH 8.0 install, received
/mnt/sysimage/var/tmp/libmng-1.0.4-1.i386.rpm cannot be opened mesg
Tested iso on CDROM ok. CDROM was created from ftp download
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Boot from diskette
2.build server with most options and all 3 GUIs
3.fails about 1/2 way thru the install process
Actual Results: Install went fine until above reported crash
Expected Results: duh..
This is a pretty new PC that has been running Win2KPro for some time. hdb was
blank and was used to install linux
Tried to install by making less option selections. Failed much earlier, in the
same way..this time on libz. Can't attach the crash file since it was
formatted for UNIX and not for DOS.
Files on CDROM that's being complained about
09/03/2002 01:27p 117,947 libmng-1.0.4-1.i386.rpm
09/03/2002 01:00p 36,660 zlib-1.1.4-4.i386.rpm
Tried installing just the RH workstation software. Failed in the same way
(zlib). The debugger said fopen /mnt/source/.discinfo not found. I presume
this is a file being written to on the hard drive being built. Fails about 1/4
way through the install process. I'm guessing this must be an anaconda bug..
File already open/locked?
At the time of failure, the subdirectory /mnt/source does not exist on hdb. I
presume that it once did otherwise the install wouldn't have gotten this far,
right? Don't see anything in the errata or bug lists about this although I
assume that it fails this way for everyone.
Found some information in a closed bug report that fixed my problem.
Apparently a lot of people are having problems with CDROMs and the update
software I downloaded served as a good workaround. As a suggestion, you might
consider implementing a less fatal error handler. You also might consider
putting this info in the errata sheets since I had to do a *LOT* of searching
to actually find teh info I needed.
Resolution == RESOLVED DUPLICATE 75377
Component == Anaconda
If you had included the Anaconda dump file when you initially filed this bug,
you would have been pointed to the bug that would have assisted you. Glad you
found the solution to your problem, though.