Re: problem with cp-command

[Peter Landgren]

Den Friday 06 July 2007 18.20.26 skrev Bob Proulx:
> Peter Landgren wrote:
> > I have done some more tests. It seems that the problem only appears when
> > I copy from one specific user.
>
> A specified user?  Or a specific file?  (I am thinking about something
> particular here...)

In the test I did I got this problem, when I copied one file from a specific 
user. The files are always copied, but the error occurs afterwords. If I use 
konqueror or krusader to copy the same file, no errors.
I have a trace of one cp-command. See attachment.

> > I created a new user copied sip-4.6.tar.gz, unpacked it and run
> > pytgon configure.py, make did a su and "make install" and now
> > everything worked just fine. So there is some problem with the "old"
> > user, but I cant' figure out what it is.
>
> Which seems to show that there is nothing intrically wrong with the
> programs or file permissions.  Is the filesystem full?
No.
address@hidden ~]$ df
Filesystem             Size   Used Avail  Used% Mounted on
/dev/sda2             9,7G   4,8G  4,4G  53% /
/dev/sda7             242G   17G  226G   7% /home
/dev/sda6             39G    5,3G   32G  15% /usr
address@hidden ~]$

> > However, I tried to install PyQt-X11-4.2 from this new user, and at the
> > end of the install process I got the same error again, when there were
> > just two files left to copy! There were hundreds of cp-commands in this
> > install.
> >
> > What's going on? Is it a problem with cp or is it a problem with the
> > Mandriva Spring distro?
>
> Do any errors show up in your system's syslog file?  (Usually
> /var/log/syslog or /var/log/messages but I am not familiar with
> Mandriva and you may have to find the right file.)  Are there any disk
> I/O errors being reported there?
I looked at both files, but found nothing (no errors) that I could connect 
with my problem.

No disk I/O errors are reported.

> This is beginning to sound like hardware failure.  It would be
> plausible to me that when certain disk sectors are accessed that it
> would generate errors and cp would fail.  But access other disk
> sections and everything succeeds.

I agree that it sounds like that. It's a new box with a 320 GB SATA drive 
using a ASUS P5VD2-X motherboard.
But it has happended last week on my old box also.

How can I test/verify this?

> > > > As I do this as root, cp should not complain?
>
> The root superuser will have permission and therefore all permission
> errors should be avoided.  But even root will fail when the filesystem
> is full.  And root has no permission over NFS.  While being root
> avoids permission problems locally it does not guarantee that there
> won't be other problems.
>
> Bob
Yes, I understand that.

I'm confused.
/Peter


-- 
Peter Landgren
Talken Hagen
671 94 Brunskog

0570-530 21
070 -635 4719
address@hidden
Skype: pgl4820

cp-problem3.txt
Description: Text document