Discuss mailing list

[Jeff Taylor <shdwdrgn AT sourpuss.net>]

For what its worth, I added testing of the T1 servers to the T2 status page 
because of this very reason... I know that NS3 (which carried bzh) has been 
failing for some time, but I've been heavily involved with an event for the past 
few weeks, so I don't know exactly how long its been failing.  With the server 
status being recorded now, we'll have a record of each server's performance, and 
can see exactly how long problems have been happening.

As for what is considered an appropriate time period... The T1 servers are 
rather critical to operations, so its important that admins keep in contact with 
each other.  I can't see any possible way that a server would just suddenly stop 
responding to all other OpenNic zones, yet continue serving the admin's own zone 
-- unless the OpenNic zones had been purposely removed.  From my point of view, 
if a T1 server is failing for more than a week and the admin has not responded 
that he is aware of the problem, or at least notified others that he was going 
on vacation, it shows a lack of responsibility.  From my own experience, I was 
in a major collision three years ago, laid up in ICU for a few days, then 
finally able to come home with a shattered leg.  Even through all that, I 
managed to have someone bring me a laptop when I got home so I could check that 
everything was still working and let everyone know I wasn't dead.  I know I'm 
anal about keeping my own systems running, but I don't think its wrong to expect 
at least periodic inspections of their servers from other admins.

And last, for the option of another admin picking up the zone... We've had TLDs 
abandoned before.  We are in a great position now to have many parts of OpenNic 
configured to automatically route around problems like this.  Every one of the 
DNS servers that slave all the zones will have a copy of the latest zone file 
for BZH, which gives us all the needed info to allow anyone else to pick it up.  
If anyone has an interest in this, by all means please speak up before the 
voting closes on the issue!


On 09/20/2011 09:45 AM, Zach Gibbens wrote:
> We did our due diligence, they metioned it themselves, they have some 
> involvement with the group trying to get it into ICANN, they saw this as a 
> way to
> Promote the idea with ICANN dragging their feet for a long time (I think I 
> recall somebody saying years, but not too sure on that)
> Their plans were for ICANN approval in the long run, so that wouldn't have 
> been much of an issue
>
> How long do we have to talk about a thread till somebody gets the message, 
> reads it and fixes it, and how long does it take to read it and say,
> whoops your right, it's broken, I'll fix it ASAP, how long has it been down, 
> over 5 days, that might be premature, I'd be thinking 7-10 (for a vote to 
> remove to start) but when the vote closes in 15 days (10 for the vote, 5 for 
> this thread) is that enough time, if we waited for 60 days, why was that 
> enough time
>
> it looks poorly on the group to have servers that don't respond, looks poorly 
> to have to change /etc/resolv.conf (or equavilant) every so often, becuase a 
> server is offline or b0rked.
> I honestly think we should have this as a policy, if your server is 
> unresponsive for one week, and an admin is unresponsive for one week, (maybe 
> a week is too long, maybe it's too short) we should have that be the immidate 
> reason to drop a server, and talk about hosting it on another tier1 or 
> dropping it.
>
> Is there a Tier 1 op that'd rather carry it than drop it?