The cost of bleeding edge?

I’ve just finished updating my workstation at home and at work to Fedora 11.  The impetuous was to cut down on the number of VMs that were running on my machine at home and reduce the number of packages I have to build for each software release.  I had a FC8 and a FC10 VM on the host which was running FC9.  I decided to shut them all down and just upgrade the host to FC11, since it had been out for a month and a half.

The upgrade itself was quite painless, but ‘fixing’ all my applications was another story.  I had to apply the patch in to get VMware Workstation 6.5.2 to run on my machine at home.

I ran into an OS bug as well.   My CentOS 5.3 VM can no longer write to the home directory that is exported from the FC11 host.  I think I actually had problems with the latest FC9 kernel because, when trying to build open-vm-tools-2009.06.18, I was getting write errors. The unusual thing is that I do not have problems writing to my home directory on my FreeBSD or Solaris 10 VM.  I believe the problem is related to Fedora bug 508095 – NFS write lock problems.

At work, I run VMware Server 1.0.9.  According to this thread,, I’ll have to rebuild the kernel.  This will have to be done for each and every kernel update.  What a PITA.

So, what have I gained by upgrading my main workstations to FC 11?  Quite a bit of pain and the voice in the back of my head second guessing my choice.  I’m now seriously thinking of reinstalling both machines with CentOS (5.3) and then just run a FC 11 VM at home.  I think both VMware applications should be supported a bit better on CentOS.  I’m a bit disappointed that 8 years after dealing with my first Linux NFS bug, on RedHat 7.1, that I’m still running across the same problem.