Lanny Marcus schrieb:
On 3/20/09, Lanny Marcus lmmailinglists@gmail.com wrote:
2009/3/20 Tru Huynh tru@centos.org:
On Fri, Mar 20, 2009 at 11:02:46AM -0500, Lanny Marcus wrote:
ae_cw_10g.databank.com 0.0% 111.6 113.6 110.7 126.5 4.8 pod22h_ae.layeredtech.com 0.0% 116.8 115.5 110.7 139.7 8.7 ??? 100.0 0.0 0.0 0.0 0.0 0.0 162.194.232.72.static.reverse.ltdomains.com 10.0% 111.4 117.4 110.9 140.3 9.8 [root@dell2400 ~]#
So, from this point of view, L3 seems to be losing >some packets.
At the end, when I ran the mtr command you used, ??? is losing 100% at one router within layered tech.com (the hop after pod22h_ae.layeredtech.com) and the next one, 162.194.232.72.static.reverse.ltdomains.com is losing10%?
162.194.232.72.static.reverse.ltdomains.com is www.centos.org
Tru: And it's losing 10%? There seems to be a larger problem, before it gets to that server, within LT. If you read the last few messages in this thread, posted this morning (GMT -5 here), by me, Rainer and Per, the LT Tech should be able to track down the problem(s) within their DC. Take care and thank you! Lanny
Tru: Until Rainer showed the command he used "mtr -c 10 -r centos.org" I was unaware of it, until this morning. Possibly the 10% loss at the server is OK?
This can have a variety of reasons - packet-filtering is not the least likely ;-) I don't think centos.org is just sitting there with no iptables etc.
I can reach centos.org from my colo-box perfectly (via lynx) - but I'm not sure why I would need 16 or 17 hops to do that - and even 20 for layeredtech.com. I can reach our own little website via 8 hops, same as yahoo.de
I should point out that I have no deeper insight into the inner workings of how internet-routing works - maybe this is all expected. But in the past, mtr was quite reliable pin-pointing severe problems.
Rainer