Hi Hans, On Wed, 2002-11-27 at 09:42, Hans Ekbrand wrote:
> Really strange, what could link these two phenomena? For a client not > to be able to boot the kernel, one might think that the kernel it got > was corrupted on some way. But why would that correspond to the uptime > on the server? There is a "trivial" in tftp, but I didn't think the > packets could be bad... > I went onsite to investigate the problem yesterday. It seems that aas more clients gets online, and network traffic increases, the problem starts to occur. Furthermore, one client that hasn't been able to boot for quite a while, was swithched with no result - then I moved that connection to another switch (from a HP 100mbit switch to a rubytech switch - also 100mbit), and it was able to boot in the first try. Most of the active LTSP client is placed on the HP switch, so I am lead to believe that something wicked happens as network traffic increases. This is q very little LTSP environment - only 17-18 active clients on a 100mbit switched LAN, so this really shouldn't be a problem. Is anyone able to comment on this ? > You have several kernels installed, why is that? The "Ok. booting the > kernel..." message indicates a problem with the kernel. Make sure the > same kernel is given the clients each time. > All clients gets the same kernel - 2.4.19-lts every time. I changed from 2.4.9-lts to see if that made any difference. > Are the clients are powered down between boots? > Sometimes yes, sometimes no. It doesn't appear to make any difference. -- Med venlig hilsen / Regards Klaus Agnoletti Junior Geek Engineer Xenux - The Linux People Bredgade 35A, 2. 1260 K�benhavn K Tel: +45 3315 8202 Fax: +45 3332 1832 http://www.xenux.dk Xenux og SSLUG har oversat teksten 'Why Open Source / Free Software - Look at the numbers' der giver seri�s argumentation for brugen af OSS. Se den p� http://www.sslug.dk/~anne/dk_why_oss_fs.html
signature.asc
Description: This is a digitally signed message part
