I have resolved the issue but want to see if other people have ran into this.
So I was in the process of replacing a 1.5TB drive with a 2.0TB drive on my server. Thought this would be straightforward, but ran into the strange message in the title. I am running 12.2 x86_64 as a headless server. I pulled out an old 500GB drive (it was the root of my former server before I rebuilt and installed 12.2, so it was only in there while I was doing setup and just never pulled it out). So I unplug that 500GB and plug in the 2TB and turn it on. After 10 minutes I was still not able to ssh into the machine, so I run around and find a monitor plug it in and am at a login prompt for the recovery console with a message about failure to load radeon/CEDAR_pfp.bin.
I find this rather perplexing, shut down the machine and make sure I have everything plugged in, and didn’t accidentally bump the videocard. Everything looks fine, I restart and after a hangup after trying to mount drives I get dumped out to the recovery console. So I end up switching the harddrives back and everything works properly, without the new drive in. Thinking that there may be something loading off of that old drive (since it was a boot/system drive of an old system) I decide to just swap it with the 1.5TB Im replacing it with. So I do this and boot up and get the same message after 5 minutes or so, now Im really confused. I’m out of drive bays but as a test I decide to leave the other drives hooked up and just plug the new drive in and see if it will boot it up with that in with all the other drives the same. Now it boots up properly.
So after that test and seeing as it would bomb out during the mount phase of boot I decide to go through and comment out the fstab entries for the 500GB drive I was removing. Removed the drive, and didnt have the issue again. Also I forgot to mention im running in a minimal setup that boots to command line, never booted in a graphical system but had to install some xWindows components to be able to install Tomcat through the default repos.
So theres my story now my question, other than has anyone else seen this behavior is why would a failed mounting of a drive cause a failure to load a gpu driver? I fully expected to get errors about not being able to mount something in fstab but did not expect it to completely bomb the entire boot process with a failure on a seemingly unrelated system.
On 2012-11-14 08:06, please try again wrote:
>
> baaldemon;2503650 Wrote:
>>
>> So theres my story now my question, other than has anyone else seen
>> this behavior is why would a failed mounting of a drive cause a failure
>> to load a gpu driver? I fully expected to get errors about not being
>> able to mount something in fstab but did not expect it to completely
>> bomb the entire boot process with a failure on a seemingly unrelated
>> system.
>
>
> I confess that I also experience such issues from time to time while
> generating incorrect fstab from script. Last example here:
> http://tinyurl.com/8be6l2n. Different Linux distros have different
> response to fstab errors, but AFAIK under openSUSE, non existing devices
> bring you to the recovery console. One might try to use the “nofail”
> option to avoid problems though.
Notice that system dumps you to recovery mode in systemd with no clear
explanation of the cause, contrary to what happened in systemv. You have
to look on certain logs, and try to guess what it is.
I have a bugzilla on this (Bug 782904) - feel free to add your
experiences, because they are not going to do anything.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” (Minas Tirith))
Yes. Last time, It took me a while, a couple misleading posts and an invalid bug report until I realized it was an fstab issue … despite I’m aware of this problem.
Thanks for the info, and for assuring me I wasnt just freaking crazy with what I experienced. I put a new comment on the bug report, but didnt have many real details to add. Now I know to make sure I update fstab before I shut down to replace a drive next time.
On 2012-11-15 02:26, baaldemon wrote:
>
> Thanks for the info, and for assuring me I wasnt just freaking crazy
> with what I experienced. I put a new comment on the bug report, but
> didnt have many real details to add. Now I know to make sure I update
> fstab before I shut down to replace a drive next time.
You elicited an interesting response, that it is kind of solved in
factory. Someone will have to test it - I can’t, yet.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” (Minas Tirith))