Bug 81653

see bug #81597 for a related comment.
note that space requirements are pretty individual, so it is hard to define some generally useful limits.
anyway, i agree that kde could check for "disk low" situations at startup, possibly by doing a ratio calculation on the df output or whatever. this is not kdm's task, though.

most important to me (!) is that the users do not get the feeling "it just logs me out again when I log in".
Then, there should come up a message "Logging you out again because no space left". That would make error tracking more easy.

ksmserver checks for disk full and some other common problems and informs the user.

Seems like that check doesn't work too well then.

I notice that if /tmp is unwritable (e.g. due to a full disk) the X server will not fully come up and startkde is never executed, so ksmserver will not have a chance to tell anything to the user.

So I suggest that kdm adopts some of the sanity checks from ksmserver/main.cpp

*** Bug 101446 has been marked as a duplicate of this bug. ***

*** Bug 138232 has been marked as a duplicate of this bug. ***

*** Bug 142775 has been marked as a duplicate of this bug. ***

*** Bug 159180 has been marked as a duplicate of this bug. ***

Hi,
If someone give me the directions to what is the first app called in startkde i can write a simple pach that:
Check for quotas and full disk space at startUp and simply do a 
KMessageBox::Warning("You runned out of space, free some files or buy a new HD :P")
then start dolphin after that check again to see if there's is at least "5 mb" free and start kde otherwise warn again...
i can also do a system service that could check periodicaly for space and quotas, if someone know if there is some working in progress or where should i do that i'll appreciate, someone told me that solid should have something like this.

*** Bug 164853 has been marked as a duplicate of this bug. ***

One should think about a warning in the logout instead of the login. 

If you try to logout while there is only 0M free, you should be warned that you will no more be able to login again.

Doesnt make much sense. Think of a multi user system. Why should a user care about the next user?

I think I have an idea how to fix this, assigning from kdm-bugs-null@kde.org to me

still exists for KDE from today

As far as I can see, kdm does not call startkde. So, yes, sanity checks in kdm are needed.

kdm's possible session types (KDE, GNOME etc) are defined in .desktop files in $KDEDIR/share/apps/kdm/sessions. There is also stated where the startkde resides that will be started by kdm. This brought me on the wrong track - kdm does call startkde.

(In reply to comment #4)
> I notice that if /tmp is unwritable (e.g. due to a full disk) the X server will
> not fully come up and startkde is never executed, so ksmserver will not have a
> chance to tell anything to the user.
> 
> So I suggest that kdm adopts some of the sanity checks from ksmserver/main.cpp
> 
wrong. startkde IS executed, I just tested this scenario.

some bugs were fixed since waldo made that comment. ;)
it is still possible that it wouldn't even get to run ksmserver - e.g., if /home is shared with /tmp and the user has a strict quota or the disk is simply *completely* full.
if you want to do it Right™,you need to create pam_diskfull (google first - maybe somebody already did :).

Thanks Oswald, but I tested in a completely-full-situation.

startkde is started, but it cannot access X (no xclock, no xmessage, ...). So my proposal is the startkde should return a special value (say 5) to indicate it found there is no more space on /tmp. kdm would need to react on this value - does this find your placet, Ossi?

bah, so the error while writing the Xauth temp file still goes undetected and the x server later refuses to allow clients to connect. this needs to be fixed.

*** Bug 191452 has been marked as a duplicate of this bug. ***

Why is this severity wishlist? In the case I encountered in https://bugs.kde.org/show_bug.cgi?id=191452 , the laptop was completely and utterly broken for my girlfriend. Thankfully, I was near so I could fix it for her.

Otherwise the laptop would have been, literally, useless to her.

If this happens while someone is abroad etc, this is a major issue!

PS: Warning at logout may help, but a warning within kdm, and a way to fix it, are needed.


Unless someone objects Real Soon, I will raise the severity to critical or major as per https://bugs.kde.org/page.cgi?id=fields.html#bug_severity

As per discussion in #kde-devel, I am setting this to grave. It will probably not help fixing anything, but at least it's tagged correctly.

http://www.kernel.org/pub/linux/libs/pam/modules.html does not list anything which could help.

I sent email to upstream.

A related bug is this: bug #205630 . 

If the disk is full the splash screen is drawn over the messagebox which indicate that the disk is full, so if the user wait for the login without a mouse click, he will never know that the disk is full because the message is hidden.

is 194806 a duplicate of this bug?

no, this report is currently about missing/incorrect error handling (see comment #20). the other one is about a workaround (quite a hack, in fact).

i'm thinking about extending disk monitor plasma applet to give a notification to user when the disk space is full

While that is helpful in general, this does not help if something else filled the disk or the user does not comprehend the warning.

Background processes, daemons or other users might have filled the disk.

*** Bug 374688 has been marked as a duplicate of this bug. ***

KDM is unmaintained and not used in KDE Plasma 5.

SDDM is the login manager used in KDE Plasma 5. If you still have this same issue with SDDM, please file an issue on the SDDM bugtracker (after doing a search for existing issues first!): https://github.com/sddm/sddm/issues/