home | list info | list archive | date index | thread index

Re: [OCLUG-Tech] Re: GCC: Request for help to get cluster booting!

  • Subject: Re: [OCLUG-Tech] Re: GCC: Request for help to get cluster booting!
  • From: Stephen Gregory <oclug [ at ] kernelpanic [ dot ] ca>
  • Date: Sun, 21 Sep 2008 21:47:01 -0400
On Sun, Sep 21, 2008 at 01:31:24AM -0400, Richard Cook wrote:
> OMG - It booted!!!!!!

excellent.

> - Rebuilt the sis900.zlilo file at ROM-O-Matic.
>   I did not remember all of the options I selected the first time
>   but I am sure I selected MORE this time and the file is smaller!

I suspect this was the key. You may just have neede a newer version of
etherboot/gPXE.

> The only problem is that the node stopped at a 'boot:' prompt until I 
> pressed Esc.  I thought I had configured it to NOT stop.  fix later.

I think TIMEOUT 0 may mean wait forever. Also be sure to specify a
DEFAULT. This is all in the pxelinux.cfg/* files. The documentation on


> Thank you, thank you, thank you!!!!

I am glad it finally worked.


> Now, if I boot another node will it conflict with the second?

It should not. NFS should not be an issue. I have 40 manchines all
using the same / (ro) and /home (rw) without issues or special
config. The only hang ups you will run into is shared /var/run /tmp
and similar places. See my earlier post about mounting a bunch of
tmpfs file systems. What I posted was debian specific. You may need to
tweek it for other distros.

-- 
sg