no X after full upgrade from test3 to upgraded pkgs for test3

Paul W. Frields paul at frields.com
Wed Oct 29 18:30:02 UTC 2003


On Wed, 2003-10-29 at 11:32, Michael K. Johnson wrote:
> > After a full upgrade this morning from older test3 to the newest
> > stuff, I'm unable to startx or to boot to runlevel 5.
> > XFREE pkgs were some of those updated and now consist of:
[...snip...]
> > I'm pretty sure there should be a bunch more from
> > rpm -qa|grep XF
> 
> You can get a log of what was on your machine before by looking
> at /var/log/rpmpkgs (or one of the rotate versions).  That can
> help you track down what went missing.

In an earlier posting I described an up2date failure where a truncated
downloaded package file was not caught by the up2date client. up2date
proceeded to perform the updates, and in the middle of one of the
earlier upgrades terminated with a fatal error. This left a number of
missing packages on the system which had been previously installed, as
apparently has happened here.

I'm not a Python programmer, but a review of the code in
/usr/share/rhn/up2date_client shows that there is some sort of RPM MD5
checking functionality included, but it apparently doesn't use the
payload. Perhaps the workflow there needs to be revisited to see if it
would be appropriate to make a more rigorous check before up2date starts
removing anything.

Again, as before, I ask that someone check my logic and understanding of
the facts or a larger situation here.

-- 
Paul W. Frields <paul at frields.com>





More information about the test mailing list