Re: RFE: skip-undownloadable - that operates when rpms aren't retrievable

Hi   to Timms

I wasn't sure from your message if you were looking for a "fix" or 
requesting improvements to a plugin 
So if I am missing the plot , I apologize up front.
the same thing happened to me today ( assuming livna ), I had downloaded 
about 423 meg
and I didn't want to clean all to download again.  so I did a clean 
dbcache  then disabled  the repo
and proceeded to update again, same thing happened due to one package 
still needing deps from the disabled repo,  I manually excluded the 3 
offending packages  ,and proceeded to update again - it did not 
re-download since "packages" were not cleaned , and all went smoothly 
except that I do have 3 packages.
I do have yum-skip-broken installed, but you are right in that it 
doesn't handle odd situations like this.

again, If this post wasn't what you wanted , I do apologize  for missing 
the boat.
thanks to all who post regularly for all the help you have provided.
Cheers,
Matt


David Timms wrote:
> Hi, it's common within the mirrors the are assigned to my region to be
> unable to download particular rpms. The error might be "byte range not
> satisfiable", mismatch, or not found etc. yum currently tries the 5 or 6
> assigned mirrors for each file, and might find that none can provide the
> rpm at this very moment. Other downloads might succeed, which would
> allow a smaller transaction set to be tested, based on the rpms that
> have been downloaded, and subtracting any whose requirements are not met.
>
> An additional message announcing the fact like:
> 101 packages set for upgrade.
> 35 packages not currently available
> 48 packages where updated. {sums didn't work because of unsatisfiable 
> deps}
>
> As far as I know skip-broken doesn't take care of that problem; the
> whole transaction will be aborted. The manual way is to either wait some
> unknown time for things to "get right" or else run yum with -x for the
> undownloadable packages.
>
> I expect that the exclusions could be automated in much the same way
> that skip-broken works, perhaps using much of it's code. I think this
> would be a good robustness enhancement to yum, that is make the absolute
> best of a poor situation {missing rpms} / partially updated mirrors}.
>
> DaveT.
> _______________________________________________
> Yum mailing list
> [email protected]
> https://lists.dulug.duke.edu/mailman/listinfo/yum
>