Hi,
I am setting up a clustter with System Imager (pre-release of v4.2, i.e.
4.1.99). The nodes are HP ProLiant SL390s G7 (half-width servers, eight
at a time in a common enclosure). We are running CentOS 6.2 (still
pretty vanilla).
So far I got the system-configurator part seemingly OK, I can get an
image, and net-boot to distribute it to a new host.
However, on the following boot (pxe: localboot) the node dies with a red
screen of death (Illegal Opcode). Just before that happens, the console
states that it will boot from local hard disk. I am not sure what happens.
The node has only a single local (internal) hard disk, so the problem
should not be related to SAN/NFS/NAS-disks, or to HP disk arrays etc.
If I change the boot order (in BIOS) on the node to be hard disk before
PXE/net, then the ndoe boots happily from the disk, so it seems like the
imaging and grub-install was done OK. (I have also been checking the
system prior the the boot, by adding a "shellout" to the installation
script). Can the PXE/Localboot be misinterpreted by the node (HP
server)? What files should I check (/etc/dhcp/dhcpd.conf and
/tftpboot/pxelinux.cfg/{default,IPHEX}
I have noted that there are a single thread on opcode problems on
sisuite-users back from 2010, and the reply from Kris Buytaert may be
related, but I am still not sure what to do. The idea of changing BIOS
(twice) to get an updated image does not really seem like a pretty solution.
All help is welcome.
Bjarne
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
sisuite-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/sisuite-users
