16 September 2000
Here we go again...
Another hurricane is brewing, this time in the Gulf of Mexico
just off of central to southern Florida. The projected path takes
it over north Florida, through Georgia and South Carolina, then
directly through central North Carolina. Though wind will not
be an issue with this storm, the amount of rain may equal that
of Hurricane Floyd of last summer which pretty well destroyed
everything east of Raleigh (and made life messy around here for
Probably the worst we will experience is a loss of power,
but the generator is in place, tested, and ready to go. We'll
keep you posted as we get closer to Monday night/Tuesday morning.
14 June 2000
Very bad. At 4:30 pm it was partly cloudy with a hint of thunder
in the western distance. At 4:45 pm, we were hit by one of the
worst thunderstorms that Raleigh has seen is decades. The thing
popped up in a matter of minutes and hammered us. We took several
direct hits by some amazing positive charged lightning bolts
which is when I literally unplugged all exterior connections
from the LAN. Then the power went out.
Though we pulled the generator in time before the UPS units
hosed, we could not reconnect the telco until 6:27; it took that
long for things to calm down. The total outage was 103 minutes
before things became stable enough to bring the whole system
back on line.
The problem is that I can protect the telco in-points from
lightening surge when we are dealing with "normal"
negative charge bolts that take off from the high points no matter
how close they are to us. We can not protect from the positive
bolts since they energize the surrounding ground as well. They
are also far more powerful than negative-charged lightening and
do a tremendous amount of damage. Fortunately, this is the first
time in the four years we have been in business that we have
had positive-charged bolts in the area. By the way...one method
of distinguishing positive bolts from negative bolts is to look
at a tree that has been hit. A negative bolt will singe the tree
and perhaps split a trunk notch if it is powerful enough. A positive
bolt makes the tree explode into flaming toothpicks.
I really hope it does not happen again any time soon.
10 May 2000
Not a good day. The primary email server went south this morning
shortly after 5 AM. It is going to take a complete repair and
restoration to get it back up and running. Now the good news...
No mail is or will be lost. Any mail that was on the machine
at the time is hosed is still there and completely recoverable
(which is what I am doing now and anticipate I will be doing
for many hours to come.) Any mail that was sent after the machine
hosed will bounce around, attempting redelivery every hour or
so until the machine comes back up. Eventhally, it will be delivered
and will not get lost.
Unfortunately, for those clients who are using both the POP
and the SMTP server that is down, you will not be able to check
mail or send mail via our system untiil the server comes back
up. I anticipate that we will be up and running by late this
afternoon, perhaps by 4 PM eastern time. Until then, I apoligize
for any inconvenience. Just be glad that we keep things backed
Email service was restored shortly before 11 AM today. Things
are now running normally, though the poor server is breaking
a major sweat handling all the backed up email. The log jam is
now cleared (1PM) and the stupid machine better behave or I will
smack it hard.
25 April 2000
Well...we played the odds and lost today. A local power transformer
on a pole just up the street blew up shortly after 5 PM this
afternoon. When things like that happen, we play it by ear as
to whether we are going to pull a generator or not. Had that
transformer actually been the one that specifically feeds our
power, I would have pulled one in and connected it since CP&L
generally takes about four to six hours to fix something like
that. But it was not on our direct feed so all I needed to do
is ride it out for 20 minutes or so with back-up power until
CP&L arrived and kicked back on the circuit breaker that
handles our feed (and which tripped when the transformer exploded.)
Unfortunately, it took them 52 minutes to do something that should
have taken five. Back-up battery dies at 27 minutes and we were
down for 25 minutes total.
I am going to have a long talk with CP&L in the morning.
3 January 2000
OK folks...this is what is happening.
Some people are having problems reaching their web sites today.
We switched backbone providers and are now on a Sprint network
with Sprint IP numbers. The change went very smoothly, but there
are about 100 domains that are in potential trouble.
And here's why...
When we register a domain name, we do so with InterNIC via
the interface provided by Network Solutions. When we register
domain names in that fashion, we are able to make changes to
the DNS records of those domains. But if a domain name is registered
directly with Network Solutions through their DotCom interface,
with register.com, or another registry where a user name and
passcode are required to make modifications, then we have absolutely
no control over making any changes.
The end result is that the person who registered the domain
originally and who has the user name and passcode must make the
changes. And here is the information you need to make those changes:
- Valerie Crisp
- 3252 Octavia Street
- Raleigh, North Carolina 27606
- NIC handle (if requested): VC198
Secondary DNS Server:
Secondary DNS Server:
Secondary DNS Server:
These changes MUST be made so that the changes take affect
by 5 pm EST on Friday, 7 January 2000. At that time, we are taking
down the BellSouth telco lines and the old class C block in its
Just as a point of note, if the Commerce Department had not
been so lax with Network Solutions in allowing NetSol to claim
ownership of the domain name registry, and had forced the new
registrars to allow ISPs to make global changes without user
names and passcodes (as was the policy since the inception of
the domain name system) then none of this would have happened.
The other folks who will be affected are those who have domains
that are something other than .com, .net, and .org. They will
have to contact the respective country host registries to make
the change. That affects about 40 domains and most of them either
have been contacted or will be shortly.
One other (hopefully) small group are those who have decided
to take over their own DNS without letting us know. People may
do so in order to manage their own email server, but want to
continue to have us host their web sites. They control their
own DNS and need to point the A name records to the altered IP
number. In short, where ever you see the 205.152.36. or 208.49.46.
class C IP numbers, you need to change them to 205.160.14. and
199.1.201. respectively. Not doing so by Friday will cause your
web site to disappear off the face of the planet until the DNS
redirect is changed.
We have tried to notify as many people as we could, but many
of the email contacts we have are outdated, phone numbers are
disconnected, and all we have is a physical address. That would
have been impossible to use to make contact.
But that is why we not only have this server status page,
we also request that people change their contact information
we have on file any time it changes. If people have not done
so or do not check this page on a regular basis, there is not
much more we can do.
For those who are caught in this switch, we are going to make
every effort to continue to contact them as we see problems arise
and apologize for any inconvenience.
Of course, feel free to contact us at any time and we will
walk you through the process if necessary.
One more thing...if your domain does not fall into those discrete
categories listed above and you still can not reach your web
site it is most likely because your dial-up ISP does not play
by the rules of the Internet. Those rules are detailed in public
documents called Requests For Comments, or RFCs. One of those
RFCs specifies the protocol of something called a Time To Live,.
or TTL. The TTL is what we set on our side for your DNS information
and tells other servers how long that any particular piece of
DNS information is good for. We have our TTLs for the old IP
numbers set for five minutes and the new IP numbers set for one
hour. That means that if you dial-in to your ISP and hit your
web page, they can hold that information for five minutes. If
you hit your page again two minutes later, the DNS information
is already at your ISP and they do not need to look it up again.
But if you hit your site after five minutes, the ISP should go
back out to InterNIC to get the updated information.
A small group of ISPs, though, refuse to abide by the RFC
governing DNS TTLs. They will cache bad information for up to
a day regardless of what we set. So if you hit your page, then
we make a change, and you hit your page an hour later, the ISP
still has the old information.
There is nothing we can do about ISPs who don't play by the
rules. You need to take that up with them and find out why they
ignore policy. In all cases I have come across, it is because
the ISP does not want to pay for the little bit of extra telco
needed to do DNS lookups and would rather cache bad information
than increase their costs by perhaps one percent.
1 January 2000
Happy New year
Return to the Server Status Menu