In a message written on Wed, Nov 26, 2003 at 05:39:33PM +0100, Jerome Fleury wrote:
Is there any relationship between this "europeanwide" above.net failure and the huge amount of DNS requests to lockup.zonelabs.com which failed that every ISP (at least in France) seem to have encountered last night ? The zonelabs.com zone is hosted on Above.net NS servers.
I replied privately, but then saw this also went to Nanog. I'm also looking into that problem. My initial observation: 1) AboveNet DNS servers were down for some people in Europe. 2) ZoneLab's other nameservers are both on the same lan and connected to our network. 3) ZoneLab's software seems to have an overly agressive retry (I've been told 2 queries per second per machine) when it can't resolve a DNS name. Clearly we can only directly address #1, which is a top priority for us, we are trying to make the customer aware of why #2 and #3 are bad. I haven't proven #3 yet, it's based on other peoples reports, so I might be wrong about the software's exact behavior. -- Leo Bicknell - bicknell@ufp.org - CCIE 3440 PGP keys at http://www.ufp.org/~bicknell/ Read TMBG List - tmbg-list-request@tmbg.org, www.tmbg.org