<div dir="ltr">Yeah, I thought it was just the opportunistic merging of the massive amount of arp traffic during multi-thousand node boot (which ipxe has changed since then), but after removing it, the table filled up from reset packets (multi-thousand servers with tons of monitoring agents that would just happen to hit at least a handful of the servers every mass boot attempt).</div>
<div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Jul 11, 2013 at 7:21 PM, H. Peter Anvin <span dir="ltr"><<a href="mailto:hpa@zytor.com" target="_blank">hpa@zytor.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">On 07/11/2013 04:20 PM, Jarrod Johnson wrote:<br>
> FYI, in the xcat branch, we don't send RST. In our case it was because<br>
> TCP services pounding randomly on IPs would cause ipxe to get neighbor<br>
> table entries preventing it from talking to intended servers.<br>
<br>
</div>That is a bit... odd.<br>
<span class="HOEnZb"><font color="#888888"><br>
-hpa<br>
<br>
<br>
</font></span></blockquote></div><br></div>