>> You forget to mention if the included test program works
>> successfully. This is an important part of making sure DazukoFS is
>> working. Please try the included test program _before_ trying any
>> other Dazuko-based software.
> Checked. The test programs runs OK.
>> Could you provide similar information about your system?
> kernel: vanilla 2.6.31
> arch : i386
> h/w :m/b ASUS P5KC, C2D E8400, 2x1G DDR2, Maxtor 500G/32M attached
> to on-board JMB363 SATA in AHCI mode
> distro: Debian 5 (lenny/stable) 32 bit
> filesystem: ext3
> video: Nvidia GF9500 + proprietary drivers v.185.18.36
> AVIRA: the latest version + latest updates (used to keep the cure CD
>> And most important, does running the test program cause the same
>> problems (when AVIRA is _not_ running)?
> Not, it doesn't.
So when the test program is running, you are able to log in, start the
web browser, etc, without any problems?
> PS: But AVIRA isn't the only troublesome piece, WEB-browsers don't
> start as well for cache in a dazuko mounted directory. And after
> reboot, Xorg crashes on attempting to start - black screen with as
> stopped mouse cursor, locked keyboard with blinking
It sounds like when you do the test program, you are only using a test
mountpoint of DazukoFS (as stated in the README). Could you mount all
the DazukoFS mounts you normally would mount and run the test program?
That should cause the same log in and web browser problems.
It would be helpful if we could reproduce the problem without AVIRA
running. If DazukoFS is mounted on /home and you run the test program,
it should trigger the same problem.
Could you also list which directories you are mounting DazukoFS onto?
$ grep dazukofs /etc/fstab
$ mount | grep dazukofs
And I assume you are indeed using DazukoFS 3.1.1:
$ dmesg | grep dazuko
> Unfortunately, me don't know now to provide the kernel crash info
> since any console i/o isn't possible when kernel crash.
The kernel crash shows the stack trace. It would be helpful if you
could at least write down the function names in that trace.
Re: [Dazuko-help] v3.1.1 causes system hangs & crashes on kernel 2.6.31 (Debian Etch)
On 2009-10-24, Bobyr Raisa Efimovna <[hidden email]> wrote:
>> I've got some "good" news. I've been able to reproduce this problem
>> with the game "armagetronad". Now I can start investigating what is
>> causing it...
> No jokes as good, just imagine if such heavy problem happens after
> incoroprating into the Linux kernel :)
If it was in the Linux kernel, we'd have a lot more help getting it
> It seems that drivers tapping into file-system (like yours ) should
> be tested on all applications which intensively write/access a lot
> of files to user homes - browsers, office packages, some games,
Agreed. That's why it is good when people like you report problems.
The problem is that the kernel is trying to write to a memory-mapped
file, even though it may only be opened read-only. DazukoFS does not
support memory-mapped writing, so does not expect the kernel to
attempt it. I will need to take a deeper look at why this is
For now, I have added an extra hook in DazukoFS to return an error
when the kernel tries to do this. It works for me. Could you also try