I have a really weird problem.

This morning, I couldn't use startx from a regular user account.

I received the error

"PAM authentication failed, cannot start X server.
Perhaps you do not have console ownership?"

I searched for the error message on google, and came up with a variety of possible causes, varying by Linux version
(I am running RH Linux 8.0).

I found a bad problem, in that I could not write to the /var/run/console directory (even as root). It said I had run out of disk space. I did a df, and it said I had plenty of space. I started thinking comething wasn't right, so I
did a:
shutdown -rF
in order to force an fsck on reboot.

Sure enough, the /var device had a serious problem, and it asked me to do a manual fsck on it. I did so, it found a massive size problem with an inode, and asked if I wanted to truncate it, which I did. It then ran an inode check, and came up with about 80K+ unattached inodes, which it connected to the /lost+found directory. Each inode had a ref count of 2, which should be 1, and fixed each one.

I finally got a healthy fsck on the /var directory, and everything rebooted cleanly. The lost+found directory seemed rather large (16384 blocks), but otherwise everything looked good.

Now, I tried again from the beginning, and got the same problem. Could not startx, same error message, everything.

I went into /var/run/console, and tried to do
echo "hello" > base, and it had the same error:

-bash: base: No space left on device

df indicates plenty of space available on /var. I tried a little later, and it worked! (what's going on here?!?!?)

Then I tried again later, and got the same error message!
-bash: base: No space left on device.

I have been pulling my hair out, and could really use some help.

What's happening.....help!

No guarantees I'll be able to come up with a fix for this, but I'll chip in my 2 cents worth.

The first thought that pops to mind is inodes. Use "df -i" to see what kind of shape you're in there.

Second thought is a bad program corrupting the filesystem. Run up2date and get all your packages up to their latest and greatest.

Third thought is that there are some ext2/3 tools for working with the filesystem, and playing with some of those might at least get you more info. (I'm assuming it's either ext2 or ext3.)

At work, our main flavor of Unix is IBM's AIX. We occasionally get messages saying "Filesystem xyz cannot allocate space. Please compact filesystem." If something is chopping your space up into little bits, it might not be able allocate a file even though there's lots of space available total.

Whether ext2/3 is vulnerable to that problem is a question for a medium deep ext2 wizard, and I'm not one; but the idea is why I suggested the ext2 tuning and cleanup tools.

Fourth thought is to look at processes, and see if you can figure out any pattern between what's running and the problem.

I could go on for a quite a while throwing out things at random, but I think I'll wait and hear what came of what I've said so far.

Good luck,

CHL

I've made some progress. It turns out I had some badblocks, now I'm trying to figure out how to get the filesystem to map around them. Here are some additional details (I switched over to the hardware forum, since I think the source of the problem is badblocks on the disk).
Here's a link:
http://www.linuxquestions.org/questi...oto=nextoldest