|
From: Ron S. <rse...@ha...> - 2005-11-19 18:45:13
|
>> nodeup : No premove function for nodeinfo
^^^^^^^^^^^^^^^^^^^^^^^
>I believe I've seen this before when the stage2 image and front-end
>had bproc mismatches, i.e. I had {re-}built bproc on the
> front-end w/o generating a new stage2 image. Could that
> be possible in your case?
I built everything 'stock' from the clustermatic rpms and even
double-checked by rebuilding the stage 2 after making any config
changes, and putting the new stage 2 onto the pxe server. I'm building
the stage 2 with the command:
[root@novac root]# beoboot -2 -i -v -o /tmp/phase2
On 11/17/05, Joshua Aune <lu...@ln...> wrote:
> On Wed, 2005-11-16 at 23:05 -0600, Ron Senykoff wrote:
>
> Do you have anything coming out of the serial console (or vga console if
> not configured with serial) on the node?
Yes... any node I try (I've tried several to make sure it wasn't bad
hardware) comes up : (.219 is the server, .220 starts the node range)
boot: Server IP address: 192.168.46.219
boot: My IP address : 192.168.26.222
boot: starting bpslave: bpslave -d -i -v 192.168.46.219 2223
bpslave: connecting to 192.168.46.219:2223
bpslave: IO daemon started; pid=3D17
bpslave: connection to 192.168.46.219:2223 up and running
bpslave: Setting node number to 2
I have now moved my install back to clustermatic 4 on Redhat 9 (hoping
that i386 builds may remedy the situation) and I get the exact same
message on the node, but I get 'Signal 4' instead of 'Signal 11' in
/var/log/clustermatic/node.X:
[root@novac root]# tail /var/log/clustermatic/node.2
vmadlib : loaded /lib/libnss_bproc.so.2 (size=3D25043;id=3D0,0;mode=
=3D100755)
vmadlib : loaded /usr/lib/libbproc.so.4.0.0
(size=3D21388;id=3D0,0;mode=3D100755)
vmadlib : loaded /usr/lib/libstdc++.so.5.0.3
(size=3D710608;id=3D0,0;mode=3D100755)
nodeup : Plugin vmadlib returned status 0 (ok)
nodeup : No premove function for nodeinfo
nodeup : Starting 1 child processes.
nodeup : Finished creating child processes.
nodeup : I/O error talking to child
nodeup : Child process for node 2 died with signal 4
nodeup : Node setup returned status 1
This was another complete from scratch rebuild. Put Redhat 9 minimal +
dev packages on laptop. Installed clustermatic. Reboot into new
kernel. Build phase 2, put that on PXE server...
Any help is much appreciated. TIA!
-Ron
|