On Tue, 2005-11-29 at 15:59 +0100, Grzegorz Nosek wrote:
> > As I said earlier in this thread, I would like to be of assistance in
> > tracking this down.
> 
> Can you set up another box as a netconsole receiver? Just have syslog
> listening on some udp port and configure netconsole on your vserver
> box accordingly (say modinfo netconsole for a brief help)
I am familiar with netconsole and have a spare box to run the syslog
host on... I'll get that up-and-running as soon as I'm home.

> > I am not that familiar with debugging this stuff, but have a box at home
> > (that has apparently just crashed as I cannot reach it atm.) on which I
> > can reproduce the hang quite easily ( tar -C /opt/data/templates/ -xf
> > TEMPLATE.tgz | tar -C /vservers/ -xf - ;# crashes my box
> > time..and..time..again). So if anyone can give me a bit more debugging
> > clue, I am willing to spend some time in tracking this down.
> >
> > The machine itself is an abit bp6 dual celeron running gentoo linux and
> > the gentoo vserver kernel from a month orso ago, which is vs2.1.0pre5
> > afaik - all prior functions crashed at the same rate as well.
> >
> 
> I'm experimenting with vservers flags now with a gut feeling that
> fakeinit might be to blame (it's messing with pids slightly, after
> all). We'll see.
On a side note, I can reproduce the bug without running any vservers
just running the below tar commands will take the box down.

On another side note ;) I remember that when running 2.4 based hosts I
experienced the same type of crashes, though less frequent, but mostly
on heavily loaded machines (lots of java threads and lots of open files)

-- 
Regards,
Dennis Roos

Network Engineer @ InTouch N.V.
Middenweg 76
1097 BS Amsterdam
Tel: +31 (0)20 6752060
Fax: +31 (0)20 6758429

-=[Assumption is the mother of all f*ckups]=-


_______________________________________________
Vserver mailing list
Vserver@list.linux-vserver.org
http://list.linux-vserver.org/mailman/listinfo/vserver

Reply via email to