>
>
>I began with a straight forward installation of Red Hat Linux 7.1 choosing
>[EXPERT] mode and "Install Everything." (What the heck? I've got plenty of
>disk space and who knows when I'll want to play with a new toy!)
>
>
Ignoring the choice of 7.1, there really is not much point to
installing everything. It's something like trying to check out every
book in a library; there's no possible way to read it all, and those
books aren't going anywhere, so why not just take what you need now? If
you hate wading through dependency issues when having to install stuff
later, grab "apt" from freshrpms.net .
>On finishing the installation I get a series of error messages flashing by
>far too fast to be read in their entirety, however they start out reading
>
>"Gtk - CRITICAL ***..."
>
I'm guessing the rest is something like "unable to open screen"; if
so, the video card likely wasn't detected or set up properly. Try
"Xconfigurator" as root and see what you come up with.
>On reboot to Linux I get a bunch of error messages like what follows. I
>have no idea what they're referring to or how many of them there are
>because (again) the text flashes past too fast for mortal man to read.
>
>"(null): The superblock could not be read or does not describe a correct
>ext2 file system. If the device is valid and it really contains an ext2
>filesystem (and not swap or ufs or something else), then the superblock is
>corrupt, and you might try running e2fsck with an alternative superblock:
> e2fsck -b 8193 <device>"
>
This is a little more odd. ext2 doesn't like having the power cut
before it's properly unmounted, since there could be some stale data in
the disk cache. The following two commands will usually recover from
that, on a partition by partition basis:
mount -o remount,ro [partition to make readonly]
fsck [partition]
What makes it odd is the superblock complaints; I've rarely seen
those, even in the above scenario. Still, give those two commands a shot.
>Eventually I get to a point where Linux flashes some sort of message about
>waiting 5 minutes because X is respawning too fast (or something similar),
>and repeats the message every 5 minutes ad infinitum. I can get to a
>system prompt with control-c and log in as root but in another 5 or 10
>minutes I'm interrupted with the same "X is respawning too fast" message.
>I can recover the prompt with control-c.
>
Red Hat uses init to start Xwindows, and init is displaying that
"respawing too fast" message. Look for a line in "/etc/inittab" that
contains "x:5:respawn:/etc/X11/prefdm", and comment it out (place a "#"
at the start of the line). Xwindows will no longer bring itself up on
boot, so those messages will disappear.
>1. What's going wrong?
>
Video card is my best guess. Maybe some odd ext2 issues, but I don't
know how the computer was shut down.
>2. How do I fix it?
>
Get a better video card? More seriously, retry Xconfigurator.
Failing that, get RedHat 7.3 and hope it autodetects things better. I
bet it will.
> Cirrus GD5440 (PCI) video card (on board)
>
Oh my. Maybe you should get one anyway.
> HDD: Primary Master (Disk 1): Seagate ST340016A (Barracuda), 40 Gbytes
> Partitions: 1st Primary - W98-DOS drive C:
> Extended partition containing 12 FAT32 logical
>partitions and then 5 ext2 logical partitions.
> 3rd Primary partition as Linux swap (125.5 Mb)
>
Twelve FAT32 partitions? I hope those existed before the install.
> HDD: Primary Slave (Disk 2): Western Digital WDC AC22500L, 2.4 Gbytes
> Partitions: 1st Primary partition in ext2 as /boot
> Extended partition containing only 1 logical partition
>in ext2 as /
> 3rd Primary partition as Linux swap (126.0 MB)
>
That's a little unusual. I'd put /boot on the first drive. Still, it
obviously works, otherwise you wouldn't have gotten this far.
>FSTAB (ORIGINAL)
>
>LABEL=/ / ext2 defaults 1 1
>LABEL=/boot /boot ext2 defaults 1 2
>/dev/hda18 /home ext2 defaults 1 2
>/dev/fd0 /mnt/floppy auto noauto,owner 0 0
>/dev/hda21 /scratch4 ext2 defaults 1 2
>/dev/hda19 /tmp ext2 defaults 1 2
>/dev/hda17 /usr ext2 defaults 1 2
>/dev/hda20 /var ext2 defaults 1 2
>/dev/hda3 swap swap defaults 0 0
>/dev/hdb3 swap swap defaults 0 0
>
I have to agree with the others, you have a few too many partitions.
/var and /tmp can be placed on the
same partition. Maybe /usr and /home as well, but you don't sound
desparate for space. As an example, one of my computers has three
partitions for Linux: swap (128MB, a bit too small), / (200MB), and
/home (3GB). /usr is redirected to /home via a symlink. /var and /tmp
still reside on / .
>Adding Swap: 128480k swap-space (priority -1)
>Adding Swap: 128984k swap-space (priority -2)
>
256MB of swap for 64MB of RAM seems a bit much; I'd go with 192MB in
total.
HJ Hornbeck
