Re: Rollbacks erasing when they should not...

[James Olin Oden <joden malachi lee k12 nc us>]

On Mon, 28 Apr 2003, James Olin Oden wrote:

> OK, I went to another system and replicated the test, and I did not get 
> the failue (I actually after that I went to yet another machine) and I 
> did not get the failure.  That is on these machines it actually erased the
> current package and installed the repackaged package, unlike the 
> appearantly broken machine where it erases the current package, and then 
> stops.  
> 
> I, also, went to the trouble of setting up the same rpm configuration 
> (i.e. used the same macro files) on both systems, and I still have only
> one machine that is broken.  I am using rpm 4.2-0.69.  Is there anything
> I could check to figure out what might be causing this most bizarre and 
> heinous behaviour?
>
I found another machine where the problem occured.  On a hunch I moved 
all the existing repackaged files into another directory.  When I did 
this, and re-ran my test, all was well.  I found which rpm is causing the 
problem, but I don't know what to do to determine what the problem is?
I can query the repackaged package just fine, and it is a fairly simple 
package to boot.  Finally, it was installed at a much earlier date than 
any of these rollbacks, so on the surface, it should not affect things.

What do I need to look for to figure out why this package is causing 
rpm to not properly roll things back?

Thanks...james