[RHCE] Red Hat 7.3 install problems.

sholton rhce@trilug.org
Fri, 14 Jun 2002 06:43:06 -0400


>i installed 7.3 on my other poweredge 1400 server and had the exact same
>scsi issue (failed to insert module during kernel boot).  after
>installation a quick 'hdparm -tT /dev/sda' yields a scsi disk throughput
>of about 20MB/s, far short of the 53-56MB/s i got under 7.2.  and you can
>tell the difference, installs that used to take 15 minutes now take over
>half an hour.  on IDE systems everything is fine.  i've had a few other
>7.3 problems as well - KDE display problems with ATI video cards (which
>most of my boxes have), major nfs problems (the daemon can take up to 10
>minutes to start, and even then it is very slow to respond, tons of
>duplicate messages in /var/log/messages), and a few other minor things as
>well.  i've gone back to 7.2 for the time being, i just don't have the
>time to troubleshoot these things right now.  but i'm frustrated because i
>didn't have to troubleshoot anything with 7.1 and 7.2 - they both worked
>flawlessly right out of the box.  i kinda feel like RH let me down a
>little with this release.
>if anyone else has experienced nfs issues with 7.3, please let me know,
>that issue really has piqued my curiosity.
>jason

I isolated my problem to a corrupt /sbin/modprobe

Here's the general instructions (since we're all here to learn) but
I don't remember the exact package name:

 # rpm -q --whatprovides /sbin/modprobe
 # rpm -V modutils

which showed /sbin/modprobe with a corrupt md5 signature.

and fixed it by reinstalling the package
 # rpm -U --replacepkgs ${install_media_path}/modutils*.i386.rpm

But that doesn't explain how it got corrupted? Now I'm interested...

However, it'll have to wait until tonight.

-- 
Innovation is a wildflower; you cannot choose where it will flower,
you can only choose where it won't.

sholton@mindspring.com