So, right now I'm in the process of reinstalling CentOS on my other computer after it totally managed to kill itself in a hardcore way. Only, I have no idea why it did it.
I haven't rebooted my system in a few days so it could've been any problem waiting to happen. But here's what I did around the time from when it seemed to be working okay to when it totally screwed up:
Long story short, I downloaded the FLA of the socket object I was using in order to debug it and figure out exactly what Flash was and was not receiving from the server and see if I couldn't find out what was wrong with it.
So... install Wine. I've installed Wine on CentOS before and have gotten Photoshop to run well under it. I've gotten Flash MX to run under Wine in Fedora, so I was pretty sure it should work. Only, CentOS's wine seems to be kinda buggy. For instance, the installation windows for most programs don't show any text at all, so you hafta kinda be familiar with how the install goes or else you'd get lost easily.
So I decided I'd "yum remove wine-*" to get rid of wine, and then install the one from Fedora 8's repo. At any rate, Fedora 8's wine had the same invisible font problems as CentOS's version, so I was like whatever and installed Flash anyway. I came to find out that Flash would crash when I attempted to open the FLA in it.
So, plan B was to install VirtualBox. I had installed vbox in CentOS before and the idea this time was to run Flash inside a virtual machine of Windows. Now, this is where things started to get a little weird.
I downloaded the RPM for VirtualBox and then clicked "Open" in my download window when it was completed. A dialog popped up saying "Unknown error." wtf? so I went to my downloads directory in Nautilus and double-clicked it. Same thing. I double-clicked an older RPM of some other program that I knew worked. Ditto.
So, I opened a root terminal and did `rpm -ivh VirtualBox*.rpm` to install it that way. It appeared to install and compiled its kernel module and everything, so it was only natural to have to restart the computer then to let the kmod kick in so that vbox would run properly. This is where things went from bad to worse.
During the shutdown process, before (or maybe after) every single line that was written, this line was also written:
Cannot open font file latarcyrheb-sun16
Also, a few of the services failed to stop properly, but that happens every now and again. All those font file errors were worrisome, though.
The computer finally shuts down and comes back up. Everything seemed fine, at first. It got to where it prints "Starting udev... [OK]" and then attempted to start the X server. It got to the part where it shows your cursor (as an "X" symbol in the middle of the screen), and then the X server crashed and returned to the text-mode boot messages. The new message:
Starting udev... [OK]
Setting keymap to en-US... [FAILED]
It then proceeded to print its normal boot messages like it would've shown in the graphical boot manager if I wanted to see the details of the boot process. Only, it continued printing the error about the font file after everything, and also virtually every system service failed to start. Example from httpd:
Starting httpd: httpd: bad user name httpd [FAILED]
It gave similar errors about a lot of different services. Finally, when the boot process came to an end, I got the X blue screen of death (the screen that you get if you manage to mess up your Xorg.conf file, for example). Only the message it had for me this time was:
The GDM user 'gdm' does not exist. Please correct GDM configuration and restart GDM.
Clicking "Ok" returned me to a fairly standard text-only mode with a login prompt. My first instinct: log on as root and try to figure out what in the world is going on.
Login failed. What?! I tried a couple more times to make sure I was typing the password correctly. I even tried logging on as my regular user account: no go. It appeared that, for some reason or another, no users existed on the system anymore--not even root.
Fortunately, this wasn't a complete disaster. I popped in a Fedora 8 LiveCD. I picked this CD because it was conveniently close by and had enough default software on it for me to access the various partitions around my computer with ease. I mounted the partitions with my CentOS install on them to recover my critical files before ending up having to reinstall the operating system. I also copied my /var/log directory, but can't find the specific log that would've captured the 50% fail rate in my bootup messages. But, I'll have time to dig through the system logs later on.
This is by far the most epic disaster I've ever experienced as a Linux user, and I can't even figure out what exactly went wrong. At least when I stupidly ran commands like `chmod -R 0644 /` I knew what I did wrong, but this was just a really large problem that appears to have sprung out of the blue. Could it have been a rootkit, perhaps? When I googled the error about the font file, the only result that matched the error exactly was on a forum where somebody replied that a rootkit or virus might've been to blame. What's the likelihood of that?