Cannot boot into openSUSE11.3 - permission problem or filesystem corruption

While trying to boot into the normal user the system aborts the login and tells me that /home does not have write permission.
The login manager gives:

could not start kstartupconfig4. Check your installation.
Error code is 3. 

If I log in as root and do a

dir

or to open some files, it works and no problem seems to exist.
All this happened for the first time after running luckybackup to fusion the old files of one home directory under EXT4 with the new files from my notebook (ext3, Opensuse 11.1 KDE3) in order to come to a unified home.
But then the system did not boot any more.
I tried to change the owner but some files did not have write access (not even as root, “the file does not exist”). So I thought I might have a a file system problem. So I thought:
unmount the /dev/sdb1 and then do a fschk on it. Runs and was perfect. So I did run it with forcing to control inode per inode. No error.
So I thought: you screwed up the system.
I then decided to reinstall OpensSuse11.3 from the scratch and create the same user for home. I formatted the rest (/swap /boot /root) and left only /home.
Installation runs brilliantly, all ok, it comes to the login…

could not start kstartupconfig4. Check your installation.Error code is 3.

So: what problem could this be. No apparent file system problem, no reaction on new install. Opens under root flawlessly.

I thought about doing ddrescue on an external disc. But then: the file system reports that I have no error. And after a new install the permission of that user shouldn’t they be fixed?

Thankful for insights on what happens here and on how to solve it.
Regards.

If you create a new user, what happens?

Check the ownership of the files in /home/yourname see if any are owned by root.

Do you have a habit of logging to the GUI as root??

@caf4926

This is the thing, creating a complete new user gives the same same error message. This is why I thought initially of the file system, not of the permissions of /home/user1.
If I create i.e. user2 it experiences the same problem.

There is a problem with files that are there (indicated in dolphin) and that are owned by user but that I cannot erase because the system responds something like: the file in question does not exist. So that would be a filesystem problem but checking fsck and also with -f parameter gives always the correct filesystem (ext4) and clean. I am puzzeled.
I do not have the “habit” of loging in as GUI - root into my home account, and the problem came up as I told you when I did do a luckybackup (that is a gui for rsync) on the home from my laptops home … which is on ext3. From that moment it was impossible to login. The disk shows still 15 GB free, so it cannot be a problem of full disk either.

Do these files have odd filenames? Or filenames with odd characters in them? ETA: If that is the case, you may wish to try placeholders, such as * or $, when using rm. (You probably know better how exactly this works than I do.)

Have you run
‘chown’ on anything?
Consider making a backup of your files, (Manually).

They do infact. The characters of the name that are “strange” - probably french accents - are in temporary files of the .kpdf folder and are marked with “?” on a black background.

I tried to run chown but the result was the same. If I do the backup, do you think I should use ddrescue or would this in your opinion perpetuate the problem? Speaking especially about permission here, since, if the filenames are not ok, ddrescue would probably not transfer the errors, right?

Then this thread may be some help on how to deal with those kind of files:
batch rename files with special characters

Correction: it should be * or ?.

OK, I will try this, but question: would this invalidate write access on a whole disk? Or maybe it would interrupt the chown command and therefore leave the messed up permissions.
I am sure that is fruit of the fact that the files were synced from a ext3 disk to a ext4 disk from 11.1 to 11.3.
Original migration on that system did not show this problem. Something was overwritten evidently. Would be nice to find out what exactly happened.
I will tell you about the result of

convmv --notest --replace -r -f iso-8859-1 -t utf8 *

On 01/30/2011 08:36 AM, stakanov wrote:

> I do not have the “habit” of loging in as GUI - root into my home
> account, and the problem came up as I told you when I did do a
> luckybackup (that is a gui for rsync) on the home from my laptops home
> … which is on ext3. From that moment it was impossible to login.

ok…let me get this right: you ran luckybackup between /home/[you]/
on a laptop and a desktop (?) machine…right?

and, what did you tell luckybackup to do? that is did you set it to

  1. make all the files on one machine mirror those on the other
    machine, which machine was to be the ‘master’

  2. backup the files on one machine to the other (not changing files on
    either machine)…

  3. try to make all home files on both machines the same…

  4. if you tried to do 4 above: did you including making the files in
    ~/.hiddenDirectories/ and .hiddenFiles the same on both machines?

  5. some other thing??

in any of the above cases:

a. did luckybackup ask for your root password? or, run from a script
with root powers?

b. the user [you] on both machines is the same string (like is it
‘exactly’ “stakanov” on both, or maybe “stakanov” on one and
“Stakanov” on the other? and, do both machines have the same UID
(maybe 1000?) for [you]

c. does luckybackup include a way for you to see the rsync commands
being used? if so, please show us the command(s) used…

d. does luckybackup have an undo option?

e. is this the first time you ran luckybackup set up to do this
operation? had it run successfully before without causing problems? if
so, what changed?

f. are both machines now broken or just one, which?

g. maybe it is called LUCKYbackup because you have to be REAL lucky
not to murder your system if you use it without understanding how to
set it up a job correctly…maybe–what do you think?

h. is using luckybackup really easier than reading man rsync–and,
asking here if you need a little help?


DenverD
CAVEAT: http://is.gd/bpoMD [posted via NNTP w/openSUSE 11.3, KDE4.5.5
“release 1”, Thunderbird3.0.11,]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

yes

  1. try to make all home files on both machines the same…

This was the thing I did.

  1. if you tried to do 4 above: did you including making the files in
    ~/.hiddenDirectories/ and .hiddenFiles the same on both machines?

No

  1. some other thing??

Hum… I would like to have a date with Keira Knightley … if you can arrange? In that case I could overlook a certain behavior of yours, some lines down.

a. did luckybackup ask for your root password? or, run from a script
with root powers?

If I would have used the “root-lucky backup” available on the menu yes. But I would have told so.

b. the user [you] on both machines is the same string (like is it
‘exactly’ “stakanov” on both, or maybe “stakanov” on one and
“Stakanov” on the other? and, do both machines have the same UID
(maybe 1000?) for [you]

The accounts are identically set up.

c. does luckybackup include a way for you to see the rsync commands
being used? if so, please show us the command(s) used…

rsync -h --progress --stats -r -tgo -p -l -D --update --exclude=**/*tmp*/ --exclude=**/*cache*/ --exclude=**/*Cache*/ --exclude=/mnt/*/** --exclude=/media/*/** --exclude=**/lost+found*/ --exclude=/var/** --exclude=/proc/** --exclude=/dev/** --exclude=/sys/** --exclude=**/*Trash*/ --exclude=**/*trash*/ /home/user /media/disk-1/

d

. does luckybackup have an undo option?

No. Luckybackup is an GUI to the rsync with a possibility to control (what I did) for the code it executes.

e. is this the first time you ran luckybackup set up to do this
operation? had it run successfully before without causing problems? if
so, what changed?

It is the first time that i tried to fusion an 11.1 home with an 11.3 home. As I did write above.

f. are both machines now broken or just one, which?

Did you read the thread before answering in this way?

g. maybe it is called LUCKYbackup because you have to be REAL lucky
not to murder your system if you use it without understanding how to
set it up a job correctly…maybe–what do you think?

You get the yellow card for this. Your quote is offensive. What about this: do you read before you write? Did you ever use the program “lucky backup” before answering in this thread? Do you thing this comment was useful, and, in prospect of future readers: for them this part of the thread is just wasted space and fuzz, don’t you think so. Please abstain.

h. is using luckybackup really easier than reading man rsync–and,
asking here if you need a little help?

Same as above applies. Thank you.

I tried one more time with chown. Everything runs fine without error. The surprise comes when trying to log into the command line as a user and from there

startx

it tells you that the errors of the x-server are not fatal…but that he cannot create /kde4. Running the login gives you “no space left on device /dev/sdb1” (which is /home). So I tried:

df -h
Filesystem
/dev/sdb1 Size 4596 Used 4416 Avail 0 100% /home

So the thing is full while having plenty of space available?

Size 4596 Used 4416 Avail 0

5% is reserved as user ie not available to

That partition is full

100% is amount used

Yes, I therefore deleted 15 GB of data. It gives the same result: full. What goes wrong?

Ok I deleted another 8Gb. So now /dev/sdb1 Size 459G Used 436G Available 0% Used 100%.
Sorry for the typo before it should have been 459 G instead of 4596

On 01/30/2011 05:36 PM, stakanov wrote:

>> 3. try to make all home files on both machines the same…
> This was the thing I did.

i have looked at the commands used by luckybackup and offer here, in
brackets following the command what each switch means:

rsync
-h [make human readable numbers]

–progress [show progress during transfer]
–stats [give some file-transfer stats]
-r [recurse into directories]
-t [preserve modification times]
-g [preserve group]
-o [preserve owner (super-user only)]
-p [preserve permissions]
-l [copy symlinks as symlinks]
-D [preserve device files (super-user only) & preserve special files]
–update [skip files that are newer on the receiver]
/home/user [origination directory]
/media/disk-1/ [destination directory]

so, you no matter what you thought luckybackup was supposed to do it
commanded, not what wanted but instead to make the files which existed
on BOTH machines to be changed on the destination machine’s
/home/you (with 11.3) to become exactly the same as those files on the
sending machine’s /home/you (with 11.1)…

except, those files on the destination machine which were newer than
on the sending machine would not be changed…

and, the files on the destination machine which didn’t exist on the
sending machine would remain as they were before the run…

and, the files on the sending machine which didn’t exist on the
destination machine would be replicated on the destination

now, to the excludes which list the directories/files on the sending
machine which should NOT be sent…and, i tell you, i have NO idea how
it came with that list:

–exclude=/tmp/
–exclude=
/cache/
–exclude=/Cache/
–exclude=/mnt/*/

–exclude=/media//**
–exclude=**/lost+found
/
–exclude=/var/**
–exclude=/proc/**
–exclude=/dev/**
–exclude=/sys/**
–exclude=/Trash/
–exclude=
/trash/

but, i can tell you that it MIGHT have resulted in a usable machine
if it had had one two more exclusions:

–exclude=.*
–exclude=./*

because in that way, all of the configuration files in your new 11.3
machine would not have been garbaged up by being overwritten the
config files in your 11.1 machine

next i kinda wonder how it used a destination of “/media/disk-1/”…was
that the same as you desktop’s /home/you ?

>
>
>> 4. if you tried to do 4 above: did you including making the files in
>> ~/.hiddenDirectories/ and .hiddenFiles the same on both machines?
> No

as far as i can tell, by looking at the command you provided that is
exactly what happened, it changed all the hidden files on the desktop
to be like those on the laptop, plus all hidden files on the laptop
which didn’t exist on the desktop before, do now…

>> 5. some other thing??
>>
> Hum… I would like to have a date with Keira Knightley … if
> you can arrange? In that case I could overlook a certain behavior of
> yours, some lines down.

sorry i can’t help you with a date, nor can help that you don’t like
my method of trying to help…

>> a. did luckybackup ask for your root password? or, run from a script
>> with root powers?
> If I would have used the “root-lucky backup” available on the menu
> yes. But I would have told so.

from what you posted, it was impossible for me to know if root had
been invoked or not…that is, how was i to know that if root was
asked for you would have said so??

> d> . does luckybackup have an undo option?
> No. Luckybackup is an GUI to the rsync with a possibility to control
> (what I did) for the code it executes.

really unfortunate it undo what it did…

>> e. is this the first time you ran luckybackup set up to do this
>> operation? had it run successfully before without causing problems? if
>> so, what changed?
> It is the first time that i tried to fusion an 11.1 home with an 11.3
> home. As I did write above.

that is an operation i would have never suggested you do…why would
you want to import old 11.1 config files into a 11.3 machine?

>> f. are both machines now broken or just one, which?
> Did you read the thread before answering in this way?

yes, i read every word of every post that had been posted to this
thread before and was available to me at the time of my posting (1210,
my time today)…and, in none of those do you mention if the laptop is
still usable–and, important question if you had actually tried to
make both machines the same (that is, all files on both machines
exactly the same, always defaulting to ending with the newest file
migrating over the oldest…and, missing files on one migrating to the
other…don’t you see now how that could be an important question?

>> g. maybe it is called LUCKYbackup because you have to be REAL lucky
>> not to murder your system if you use it without understanding how to
>> set it up a job correctly…maybe–what do you think?
> You get the yellow card for this. Your quote is offensive.

i’ll take the yellow…what i said didn’t help…

but, the fact is (i think) using luckybackup you have killed your 11.3…

> What about this: do you read before you write?

i do and i did…

> Did you ever use the program “lucky
> backup” before answering in this thread?

no, but since you told me luckybackup is only a GUI for rsync i knew
all i needed to know about what it was capable of doing…which is why
i asked the questions i did…

> Do you thing this comment was
> useful, and, in prospect of future readers: for them this part of the
> thread is just wasted space and fuzz, don’t you think so. Please
> abstain.

well, no it is not useful, what i should have said was: i can’t
imagine that leaning how to use luckybackup is easier than learning
how to use rsync (which is pretty easy)…

>> h. is using luckybackup really easier than reading man rsync–and,
>> asking here if you need a little help?
> Same as above applies. Thank you.

unfortunately, i can only guess the damage would have also been done
had you used the command line, unless you knew not to overwrite the
config files on the newer system…

i remain VERY very confused about how you can add a new user and also
have the same problem unless, somehow the rsync command changed
system files outside of your home–and, again i mention that i do
not understand the command’s excludes (which seem to be mostly aimed
at not moving system level files, which shouldn’t even be looked at
when syncing home to home!!

i wonder how you added a new user?

well, i also wonder who is the owner and who can read/write
/home/[you]/.Xauthority and /home/[newUser]/.Xauthority

good luck, but i imagine you have a situation there which is gonna be
easiest fix by a backup of data, format and install and do NOT try to
fusion in an old system…instead, copy only the data, never the
configs…


DenverD
CAVEAT: http://is.gd/bpoMD [posted via NNTP w/openSUSE 11.3, KDE4.5.5
“release 1”, Thunderbird3.0.11,]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

Did you use the rm command to remove the files? Or did you use the GUI?
If GUI the the files are still there in the trash. Did you empty the trash?

Did use the rm command. Did control for the trash. The file system was compromised. So I did good primitive:

mkdir /mnt/temp
mount /dev/sdbc6 /mnt/tmp
cp -ax /home /mnt/tmp

Once done I did format the whole disks (sda, sdb) with a fresh install.
Now 11.3 runs well. I will afterwards copy the whole thing back. And I expect no problems.
Will tell you whether all works. I suspect that has nothing to do with attributes but with the file system permissions of the Ext4. As you can see above, when I did erase GB of data (with rm from the command line) the “available space” was left to 0 but the used space reduced corresponding to the erased data).
Root was able on the command line to mount and dir the data, and to do whatsoever operation. By the used space the disk should have had more then 15GB free and available.