Testing a theory – Attempting to troubleshoot a Fedora 21 install
Nothing is more frustrating, than a researcher, programmer, tester, or any other similar position attempting to document a bug, and when attempting to re-create the situation, you are unable to replicate. Testing a theory about a bad installer in Fedora 21 today, and just my luck, I was unable to reproduce the problem I was encountering on my desktop, in my virtual environment.
I spent this last weekend, attempting to install/re-install the latest Fedora Linux release. I have already backed up all my data, and done everything I need to do, in order to prep for the re-install. I figure my situation is not super unique, but probably a little more complicated than the average users’ install.
So, to get to details, I have 3 HDs in my desktop computer, I have 1 256G SDD (labeled sdc), and 2 – 3TB drives (labeled sda and sdb — Combined we’ll call md0). What I like to do with my install, is to keep the OS, and system specific binaries on the SDD and keep any data that I want on the 2 HDs; which are in a LVM – Raid 1 – LUKS (encrypted) setup. With this setup, it gives me approx. 3.1 TB of storage, which is good enough for my desktop.
So, to go into some more details, here is the basic layout idea (not exact, but just a general idea):
/ - sdc /boot - sdc /boot/efi - sdc /home - md0 /var - md0 /opt - md0 swap - md0
When attempting to install, and specifically at the partitioning section of the install, I kept running into issues. All being said, it took me approximately 7 attempts to mix up the order, of how I create the partitions, before the install finally accepted it, and started installing it.
I noted the order in which the install finally worked, and documented every step, of every failure and success in my head, so I could document it later.
Tonight, deciding to try and be helpful to the linux community, and to potentially assist the Fedora distribution, I thought I should document the failure, how I got it there, and take screenshots of every action along the way. Testing the theory that caused me to fail on my desktop install, I created a near replica setup in VirtualBox, and attempted to recreate the issue.
Again, taking all the time to screenshot everything along the way, I was unable to produce the same results that had occurred to me on my desktop. Figuring it must be a fluke, I attempt one last time, to get a failed install in VirtualBox, and once again, I am unable to replicate the results that I had on my desktop system.
I wish there was a positive moral to my story of testing a theory, but unfortunately, I have none at this point. Attempting to re-create the issue and heavily document it, resulted in a fair amount of lost time; but my thoughts were in the right place. Trying to contribute to the community and the Fedora Project.
Here goes to the next theory I attempt to test, and work with, and hopefully it will heed the proper results.