SOLVED - KDE startup problems after installing kdevelop

ogozhan, we need more information. First, when your system first boots up the partitions will be read only. That's to be on the safe side. Did you let the thing boot up entirely or panic and restart the system. If so, that's why it then wanted you to do an fsck on the disk(s).

More important than the above though: Please provide us with some specific information on your computer hardward (brand, model, cpu, amount of RAM, size of hard drive(s)) as well as what version of Mepis you are trying to use. Also, it sounds like you have actually installed Mepis to your hard drive, but could you please confirm that, or are you trying to run from the Live CD?

Also, before these problems happened, and before you tried to load Kdevelop, had you successfully installed any other software on your system (besides the main OS)? Because I'm wondering what messages you saw during the load with synaptic. Did you read the messages from synaptic? Were you informed that "such and such packages will be removed"?

Please give us a little more info to go on and we'll see if we can help you clean up the mess

Jon

All right, I use a Celeron-1.7 with 256M ram. 40 GB Samsung HD. Mepis 3.4-3 on my hard-drive. Synaptic was working well. I did install some software before this with synaptic. It did not complain anything before. It just suddenly stopped the last time I used it and complained about the read-only filesystem. After I reboot I did see some errors like:

mtrr: base(0xe0020000) is not aligned on a size(0x180000) boundary
hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=73047425, sector=73047425
ide: failed opcode was: unknown
end_request: I/O error, dev hda, sector 73047425
...
EXT3-fs error (device hda3): ext3_find_entry: reading directory #583447 offset 0
Aborting journal on device hda3.
ext3_abort called.
EXT3-fs error (device hda3): ext3_journal_start_sb: Detected aborted journal
Remounting filesystem read-only
mtrr: base(0xe0020000) is not aligned on a size(0x180000) boundary
ext3_abort called.

I investigated them and they led me to a result that I had bad sectors on my partition. I made a "fsck -cc /dev/hda3" but it reported no bad sectors or something like that.

I decided to reinstall Mepis 3.4-3 on another partition (/dev/hda6) with reiserfs. I was happy at the beginning. I installed anjuta with synaptic again. After the second or third reboot, this time kdm did not start, waiting with a black screen and a flashing cursor at the top left corner.

I decided to login from a console and then do "init 3; startx" from there. It worked. Now I write you this message from that kind of session. I changed my default runlevel to 3 in /etc/inittab.

I remember when using Mepis 3.3 the same thing happened and I decided to install Mepis 3.4.

Any advices are welcome. Thank you in advance.

Hi oguz. I'm wondering if your hard drive is getting ready to give up the ghost! Do you have another hard drive that you can test an install onto? Because if the drive itself is starting to go bad then creating new partitions will only delay the inevitable

Jon

Jon, I can start X session with startx and can work without any problems But I realized the problem is kdm does not start up. Actually I like doing "poweroff -hfi" to shutdown my computer and I think the problem can be related with that behavior.

Also I wonder if mepis switched to some sort of quick boot up becuse the same problem did happen with my Mepis 3.3 system once when I did play with such a quick boot system so much.

Can you get into kde as user root? In other words, does kdm actually start (with its login screen asking for your name and password), can can you then log in as root? What I'm wondering is if you could have set up and autologin for a particular user and that user's kde setup is now messed up.

A quick test would be this. Let's say your username is "foo". Go to the /home directory as user root and rename the foo directory to foo.old. (mv foo foo.old). Then, make a new foo directory with nothing in it. Try restarting X (ctrl+alt+backspace). If user foo is the autostarting user, and there is no kde "stuff" in /home/foo, then it will be recreated. It may not be perfect, but will kde then start up? If it does, then the problem isn't with kdm/kde, it is with the account.

That's one thing to check. Assuming that you can actually get into kde via user root, see if you can get into the kde Control Center, then System Administration, then Login Manager. Get into Administrator mode (you should be already since you were able to log in as root). Go to the Convenience tab and see if "Enable auto-login" has been applied, and with what User name. I am using an older version of KDE (3.3.2) so I'm guessing some of my info is dated. If you cannot start kde at all (here's the dated info) you may be able to correct it from the command line. In rev 3.3 of kde much of the configuration info is kept in the /etc/kde3 directory. The file kdmrc has much of the startup stuff. Save a copy of the file and then examine it for the startup options and possibly change them. If you make any changes, restart X again and see what happens.

Let me know if any of these stabs in the dark caused anyone to scream

Jon

excuse me I have told you incompletely. The story is: KDE starts and does not make a problem when I login from command prompt with an ordinary username (i.e. foo) and then do "startx". When I do "startx", KDE is automatically choosen and starts up. But when I try to run KDM, it does not. Neither when I try from command prompt nor I set default initlevel to 5. KDE works good. KDM is lost. There are no autostart accounts defined.

I will examine kdmrc. I hope it works.

Thanks.

oguz,

Here is more info from one of my systems. I just did a check to see what's running, what scripts are used, and when (run level) those scripts are executed. Below are the results. Compare them to your system.

$ ps -ef|grep kdm
root      4098     1  0 May01 ?        00:00:00 /usr/bin/kdm
root     11792 11761  0 May02 ?        00:00:13 /usr/bin/kdm_greet
$ ll /etc/init.d/*kdm*
-rwxr-xr-x 1 root root 3363 Sep 24  2004 /etc/init.d/kdm
$ ll /etc/rc*d/*kdm
lrwxrwxrwx 1 root root 13 Oct 19  2005 /etc/rc0.d/K01kdm -> ../init.d/kdm
lrwxrwxrwx 1 root root 13 Oct 19  2005 /etc/rc1.d/K99kdm -> ../init.d/kdm
lrwxrwxrwx 1 root root 13 Oct 19  2005 /etc/rc4.d/S99kdm -> ../init.d/kdm
lrwxrwxrwx 1 root root 13 Oct 19  2005 /etc/rc5.d/S99kdm -> ../init.d/kdm
lrwxrwxrwx 1 root root 13 Oct 19  2005 /etc/rc6.d/K01kdm -> ../init.d/kdm
$

Jon

Examining /etc/kde3/kdm/ directory led me to do
ln -s /usr/bin/X /usr/X11R6/bin/X
and kdm turned back.

Thank you Jon.

Note: I remember when I upgraded some program (probably kdevelop) with synaptic. It adviced me to remove /usr/X11R6/bin/. It must be the reason for this problem.

oguz, Great that you found that link and reporting back to us rather than just saying "solved" with no supporting solution.

I just posted a lengthy thread on one of the things that I think may be related. It is at:
http://www.mepis.org/node/9853

I will check your information against my "trouble boxes" and see if it might be related.

Thanks again for getting back to us

Jon

oguz, your information lead me to try a similar thing on my servers. It turns out that you have found the solution to both of our problems

You may want to "generalize" your symbolic link just a bit. What I ended up doing on my systems was

# ln -s ../X11R6/bin /usr/bin/X11

This points the system to the entire X11 command directory rather than to a single command.

But after adding the change above, I was able to load the xfree86-common package with no problem.

So now we know the "problem package": xfree86-common
and the "problem solution": "ln -s ../X11R6/bin /usr/bin/X11"

Isn't collaboration to solve problems great!

Thanks again.

Jon

I did change my harddisk a while ago because it was a faulty one with bad sectors on it that could not be repaired. This can possibly be the cause of the chain of problems begin.