xen 2.0.3 self reboot machine

xen 2.0.3 self reboot machine

[Dev Fideld]

Hello,

I installed Xen 2.0.3 on a P2-233 with 224MB.

I boot on a modified 2.6.10-xen0 kernel: I added support for
netfilter, natsemi &
pegasus network driver, uhci-hcd, usb-storage, sr_mod, sd_mod.
dom0 is Debian Sarge.

I started xend which warns about an API mismatch but starts fine.

I can start a dom1 with xm but after several minutes the machine
reboots just like
if I'd pressed reset button (ie it runs fsck on all partitions,
rebuild raid unit...). I have
 nothing in logfile in dom0 nor in dom1.

It can live 5 minutes or up to 20 but always reboot at some time. Even
if I don't hit a key or I unplug network cables.
I tryed with dom1=adamantix or dom1=woody or dom1=slackware without luck.

I noticed something strange about memory:
machine has 224MB (corectly reported by my older kernel)
I have "kernel /xen.gz dom0_mem=180224" in grub so dom0 has 176MB (171
reported by xm list)
When I run a dom0 configured with "memory=24" xm list reports 24MB but
xm info reports
memory                 : 224
free_memory            : 11

Where are the 13MB missing?

I'll try a kernel 2.4 and xen-testing or xen-unstable to see if it helps.
Do you have any hint?
If you need more informations, please ask.

Dev.

PS:


dom1 config:

name = "justatry"
kernel = "/boot/vmlinuz-2.6.10-xenU"

memory = 24
disk = [ 'phy:mapper/raid-try,sda1,w', 'phy:mapper/raid-tryswap,sda2,w' ]
root = "/dev/sda1 ro"

nics = 1
vif = ['mac=00:50:76:cc:bb:aa, bridge=br0']

~


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

Re: xen 2.0.3 self reboot machine

[Barry Flanagan]

Dev Fideld wrote:

> Hello,
> 
> I installed Xen 2.0.3 on a P2-233 with 224MB.
> 
> I boot on a modified 2.6.10-xen0 kernel: I added support for
> netfilter, natsemi &
> pegasus network driver, uhci-hcd, usb-storage, sr_mod, sd_mod.
> dom0 is Debian Sarge.
> 

I had this exact same problem. Also with Debian Sarge and a recompiled
kernel. The console showed a kernel panic complaining about a lost
interrupt.

I deleted the entire source tree, and started over. This time the ONLY
change I made to the kernel config was to add the netfilter stuff. I didn't
touch anything else.  I think I did have some usb previously, but none of
the other settings you mention. The server is now up over 24 hours and
seems to be fine.

Hope this helps.
-- 

-Barry Flanagan


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

Re: xen 2.0.3 self reboot machine

[Dev Fideld]

Hello Barry,

> I deleted the entire source tree, and started over. This time the ONLY
> change I made to the kernel config was to add the netfilter stuff. I didn't
> touch anything else.  I think I did have some usb previously, but none of
> the other settings you mention. The server is now up over 24 hours and
> seems to be fine.

I switched to xen-testing and kept only netfilter and natsemi modules
in my own kernel-2.6.10-xen0 kernel.
System has been alive for 2 hours so there might be a problem in usb part.

I'll investigate deeply on Monday.

Cheers,
Dev


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

RE: xen 2.0.3 self reboot machine

[Ian Pratt]

> I can start a dom1 with xm but after several minutes the machine
> reboots just like
> if I'd pressed reset button (ie it runs fsck on all partitions,
> rebuild raid unit...). I have
>  nothing in logfile in dom0 nor in dom1.

Leave the dom0 in console text mode rather than starting X and see if
you get an Oops message prior to the reboot. Alternatively, connect a
serial console to the machine as this will make it easier to capture any
message that comes out.

Bulding xen with debug=y may give you more information.

Best,
Ian


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl