Richard, I sorted out the problem, there seemed to be an invalid character in the DHCP record, that was the first problem, the second was that the LUI config script I was using under webmin was causing a number of errors.
The first was when creating the hosts file it was muddling up the IP address for the nodes, instead of formatting it to "192.168.2.2" it would format it as "19268.2.2.1". That error I picked up and just entered manually. The next "STUPID" mistake was that it was naming the kernel image a different name to what it was looking for, easily fixed. This would make the machine try and boot but seemed to time out and go into sleep mode, I have no idea as to why or how this one fixed itself but it may have been when I deleted a line from the end of the DHCP record that looked a little unusual. Then the node would not partition, Both machines are identical and both HDD's are the same size, the LUI config script had interrogated the Head Node and found out what the partitions were then tried to use these to setup the node, BUT... there was a huge decimal place number that seemed to be causing problems when creating the partitions it was something like "18439.39521034578561293". I stripped off the decimal place and rounded the partitions then tried again and it worked perfectly. I'm not sure that its all working correctly YET as when I submit any command to it from the C3 tool it takes ages to get a response. I'm going to do a reinstall of the entire thing including the head node and this time I will use the DHCP server that is running on my router to assign the addresses rather than have it on the head node. Thanks for the assistance. Regards Andrew Hooper ----- Original Message ----- From: "Richard Ferri" <[EMAIL PROTECTED]> To: "Andrew Hooper" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Thursday, March 28, 2002 1:48 AM Subject: Re: [Oscar-users] Install Problems > Andrew, > It looks like tftp cannot find the boot kernel. The way this works is > that the client broadcasts on the local LAN looking for a dhcp server. > The DCHP server (192.168.1.1) responds with the IP address for the > kernel server and the name of the kernel. It should respond with a > server ip address of 192.168.2.1, and a boot kernel of > /tftpboot/vmlinuz-2.4.6-1.msc-up. Then the client should tftp the > kernel over the network from the kernel server, read it into memory and > boot the client using this network boot kernel. > > Now, is the kernel actually in /tftpboot on the kernel server > (ESC00)? Assuming that the kernel is actually in the right place, with > the right name, I would suspect that either dhcp or tftp is failing. > > You can try to tftp a file from ESC00 to ESC00 (tftp to itself) by just > typing in the tftp command, and giving the ESC00 hostname. This will > tell you if tftp is happy. > > On your dchp server, you might want to kill dhcp and restart it with > the -d flag, for more debug information, as in: > > /usr/sbin/dhcpd -d5& > > Also, with dhcp errors, you should see the cause in the /var/log > messages file. > > After you take a look at tftp and the dhcp messages, please post again. > > best, Rich > > Richard Ferri > IBM Linux Technology Center > [EMAIL PROTECTED] > 845.433.7920 > > On Tue, 2002-03-26 at 22:04, Andrew Hooper wrote: > > Im new the the concept of clusters and remote boot networking etc. > > > > I have 2 1.1G machines that are identical, listed below is the config > > of my network and the comnputer that will opperate as a primary > > machine for the cluster. > > > > I have a firewall that assigns addresses via DHCP, this is a linux > > machine and serves the rest of my network. > > IP of this is 192.168.1.1 ES1 > > > > > > The machine I will use for the server in the cluster is > > Name: ES00 > > eth0 requests an IP from ES1 and can communicate with no problem > > eth1 has an IP address of 192.168.2.1 > > > > I setup the LUI script to use ESC00 as the server and assigned that > > the IP 192.168.2.1 > > > > HOSTS > > 127.0.0.0 localhost > > 192.168.2.1 ESC00 > > 192.168.2.2 ESC01 > > > > HDD Partitions are as follows. > > / > > /boot > > /home > > swap > > > > NFS Exports the following > > /tftpboot > > /usr > > > > I run LUI via Webmin and the relevent files seem to be > > created with no errors. A message stating that i should > > insert the boot disk into ESC01. > > > > I boot ESC01 and the network card is found, IP gets assigned > > and I receive the error message... > > TFTP Error 1 (File Not Found), Unable to load the file. > > > > Searching for server (DHCP).... > > <sleep> > > I have managed to get past this once, I got the mesage > > Me: 192.168.2.2, Server: 192.168.2.1 > > Loading /tftpboot/vmlinuz-2.4.6-1.msc-up <sleep> > > <sleep> > > <sleep> > > And it repeated this for ages then muttered something about > > file not found.. > > > > I would appreciate any assistance. > > > > Regards > > Andrew. > > > > BTW: Im using MSC.Linux > > > > > > > > _______________________________________________ > > Oscar-users mailing list > > [EMAIL PROTECTED] > > https://lists.sourceforge.net/lists/listinfo/oscar-users > > > > _________________________________________________________ > Do You Yahoo!? > Get your free @yahoo.com address at http://mail.yahoo.com _______________________________________________ Oscar-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/oscar-users
