Redhat cluster suite stops FAI PXEBOOT after network up

[Chris Jewell]

Hi all,

Just a quick gotcha, I think.   I suddenly had an issue reinstalling a FAI (Ubuntu 10.04 stock version) client in my cluster the other day.  The client would boot from the tftp server in the normal way, retrieve its kernel and initrd, and then stop abruptly after the network interface was brought up by the initrd scripts.  This *looked* like the old problem that occurred with multiple network interfaces.  However, after putting some judicial 'echo' statements in the initrd scripts, I determined that this was not the case.

What I found was that having experimented with GFS2, I had the redhat cluster suite running on one node (I had removed it from the others).  It was creating a network storm on port 5404 and, for some bizarre reason, this was interfering with my FAI client installation.  Once I had shut down the offending daemon on the affected node, I was able to resume normal service.

I presume that this is a bug to do with the Linux implementation of NFS and/or tftp, and not just FAI, but hopefully my observation may help someone in future!

Cheers,

Chris


--
Dr Chris Jewell
Department of Statistics
University of Warwick
Coventry
CV4 7AL
UK
Tel: +44 (0)24 7615 0778