We have been following the email list discussion regarding recent problems with yum and the upgrade to the most current CentOS 7. We have also tried some of the remedies suggested or recommended. None of these measures has yielded success, and it is not totally clear whether or not our most recent yum issues have the same root cause as those experienced by others.
Fortunately, our yum execution after the most recent updates came available was on one of our virtual machines. This is a CentOS 7 system running on VirtualBox and hosted on a Windows 7 system. It is typically used for verifying system updates and applications before they are introduced to regular Intel-based systems. This virtual CentOS 7 system has also been faithfully updated for over a year by using yum and accepting all updates when they become available.
The truly remarkable thing is that yum does not finish with any identifiable errors. In fact it does not finish at all and seems to be in a continuous loop of some kind that has run up to ten minutes, so far. Information scrolling by seems to repeat every few seconds with the names of a set of packages repeating over and over. Many of the messages indicate package conflicts or a dependency problem of some kind.
This "loop" never allows progress to the point where one would answer yes or no the to the updates. With no end in sight, we chose to kill the yum process. Shut down and restart of the current CentOS system yields a system that appears to work well with no issues related to killing the yum process. Four separate yum update sessions have been attempted, and all have been killed.
All data has been backed up properly and our application reload plan is in place, with the exception of some from epel which have caused previous problems, that others have noticed recently also. Our current plan is to remove the current CentOS 7 virtual system and install the newest CentOS 7 available. We have no plans to run yum update on any other CentOS 7 systems.