Discuss mailing list

[Jeff Taylor <shdwdrgn AT sourpuss.net>]

The digs are performed from NS0, however until yesterday I was also running the 
same reports from my own server and comparing results...

Generally speaking, the results from both hosts matched up.  Queries to the T2 
servers either went smoothly, or there were obvious problems getting reliable 
answers.  During periods when there was a problem with the host servers (NS0 or 
my server), the problems were consistent across ALL T2 server results, usually 
showing as either all yellow or all red.  If the host testing is running 
smoothly, then you will see the type of scattered results that are showing up 
now.  Note that most of the T2 servers respond consistently and are green all 
the way across.  Some servers might have one or two yellow dots which may 
indicate they were receiving high traffic loads at the time.  If you take a look 
at ns4.de and ns5.de, these are ipv4 and ipv6 ports to the same server, and the 
periods of yellow warnings match up almost identically, showing that the problem 
is with the server or its local network.  However in the case of ns6.fr and 
ns7.fr, all of the failures are showing on the ipv6 port.  To me this would 
indicate a problem with the IPv6 gateway, and not a problem with the server or 
its network connection.

If you want to see an example of host-side problems, take a look at my results 
for today between 0100 and 0830.  You will note empty bands where results could 
not be collected, and areas of yellow or red.  These stripes across multiple 
servers at one time indicate networking problems on the host... in this case my 
server was performing some intensive maintenance that hogs all CPU cycles.  (As 
a side note, my server is showing little or no yellow areas because I am now 
only performing minimal testing... does the T2 respond with a root zone or 
not... and so the results no longer match the full testing being performed from 
NS0.)

I have considered running status testing from multiple servers, however over the 
past few months that I have been setting up this code I have noticed that it is 
easy to tell when there is a problem with the host, and if you ignore results 
within those periods, the rest of the chart gives a good indication of how 
reliable query results are for each T2.  More importantly, having the chart with 
viewable history allows each T2 operator to see if there are patterns to any 
problems - such as bad response time from their machine every night during a 
maintenance cycle - and they can make adjustments based on that information.


On 08/15/2011 11:27 AM, Hospedaje Web y Servidores Dedicados wrote:
> And the digs are done from a single server? or multiple, maybe to add the ip 
> to whitelist
>
> thanks.
>
> Ing. Alejandro M.
> Hospedaje Web y Servidores Dedicados
> http://www.dedicados.com.mx
> ------
> correo / msn: ventas AT dedicados.com.mx
> skype: dedicados
> ------