edu.ph outages and bombim cadiz’ post on pinoyblog.com
March 1, 2005I’ve just noticed this
post on pinoyblog.com from Bombim Cadiz , who runs the EDU.PH domain.
It’s in response to an open letter
we made to Ver Pena of the CICT.
Here is Bombim Cadiz� post:
*
* Now that you have put your technical eggs on the accuracy of this
*”study” to disparage PHNET, let me just point that this study is not
*accurate.
*I didn’t want to bring it up yet so that you would dig yourself into a
*deeper hole but because I think you have already hit rock-bottom, it
is
*time
*to point out the mistakes.
*
*Not having any details about the methodology of the study, I will just
*point
*out that the study is inaccurate.
*
* Let me just point out the differences in the downtimes of EDU1
*(gomez.ph.net) and EDU3 (dns.ph.net):
*
*(1) On Dec1, EDU1 was shown to be down for 5:15 hours but *not* EDU3 –
a
*difference of 5:15 hours.
*
*(2) On Dec 5, EDU1 5:21 and EDU3 5:15 — a difference of 6 minutes.
*
*(3) On Dec 6, EDU1 was down for 17:04 and EDU3 for 16:29 — a
difference
*of
*35 minutes.
*
*(4) On Dec 7, EDU1 down for 9:32 and EDU3 down for 8:29, a difference
of
*1hr and 03 minutes
*
* It might interest you to know that gomez.ph.net and dns.ph.net are
*ONE
*AND THE SAME MACHINES WITH THE SAME IP Address of 165.220.1.1)*.
Therefore,
*their downtimes should exactly match.
*Because your “study” has data points which show that they do not match,
*your “study” can not be trusted.
Ouch! I’ll ignore the vitriolic language for now and focus on the
technical data.
(4) here is the Dec 1 graph for EDU1 - the DNS server at gomez.ph.net:
And here is the graph for EDU 3 - the website at http://dns.ph.net.
Bombim,
You’ll notice that both services were indeed down until 845 am.
(Presumably the server was rebooted at that time? Perhaps you can upload
your logs - as promised - so we can verify the cause of the outage?)
Notice also that the DNS server at gomez.ph.net experienced intermittent
outages from 4pm-615pm that same day while the web server at dns.ph.net
began to respond rather sluggishly at the exact same time: it was taking
20-96 seconds to respond to http GET queries during that time frame. Our
logs don’t show the web server as being down because DNS uses UDP and is
more sensitive to timeouts. HTTP, on the other hand, is a session based
protocol and has longer timeouts. HENCE the discrepancy in outage times.
(1) here is the Dec 1 graph for EDU1 - the DNS server at gomez.ph.net:
And here is the graph for EDU 3 - the website at http://dns.ph.net.
You’re absolutely right, EDU3 was down as well that day - for 2 hours and
25 mins! We just didn’t include it in our open letter as we only wanted
to point out the egregious outages (eg - More than 4 hours). (We didn’t
have room for a full page ad 
)
Notice however, that you had DNS outages (on EDU1) at 1 am, 3am-6am,
1030am -1145am, 630pm-12 midnight. Observe that at those times your
webserver started to respond sluggishly. Likewise, this explains why
EDU1 and EDU3 may be the same server, yet the outage times are different.
Perhaps another cause of your problem is that you have a slow upstream
link that tends to get congested?
(3) here is the Dec 6 graph for EDU1 - the DNS server at gomez.ph.net:
And here is the graph for EDU 3 - the website at http://dns.ph.net.
The outage times are practically identical. The most likely reason for
the 35 min discrepancy is that we don’t probe your machines every second
of the day. We only send out a probe once every ten minutes. If there is
an outage, then we send out the next probe (from a different location on
the Internet) a few seconds later - just to confirm the problem. The
probe for EDU1 does not go out at the same time as EDU3, and hence you
can’t expect the outage times to be exactly identical.
I hope that answers your questions.
The key point we are trying to make is that it’s TOTALLY UNACCEPTABLE that
EDU.PH nameservers and Websites are down this frequently. It should be
fixed AS SOON AS POSSIBLE!
The other point we’re making is that the outages on GOV.PH are even worse
than EDU.PH. We just do not understand how Mr. Ver Pena (of the CICT)
can, after working on the DotPH guidelines for 18 months, be totally
ignorant of the problems existing in Gov.PH and EDU.PH, AND do nothing
to fix these problems, YET to attempt to shut down the one Registry -
DotPH - that is operating properly!
I think the CICT should really focus less on Politics, and more on Technology!
Previous Comments
It seems that as soon as Mr. Cadiz points out the discrepancies in Disini’s data, Disini just comes up with a facetious excuse to cover up the discrepancy!
Mr. Disini should focus on the real issues instead of making up data to disparage his critics. Such ad hominem attacks do not change the fact that he treats a public resource like private property. He has misappropriated this resource and yet is accountable to none of the stakeholders in Philippine community. He should be replaced.
Posted by Maddog at August 3, 2007, 8:37 pmYES!
Posted by JOHN JOHN at May 24, 2010, 12:02 amMaddog you’re awesome! Disini amassed wealth like……
Posted by fer at May 27, 2010, 12:30 am
I looked at the heading and saw “EDU.PH outages” and did not bother to
read the entire thing anymore. The heading is already wrong.
For the sake of argument, let us assume that EDU1 was down forever. It
would still leave two other EDU.PH DNS servers functioning. Thus, even if
your “study” was correct, it would only prove that *some* of EDU.PH’s DNS
servers were down *some* of the time.
To prove that there was indeed an outage of the EDU.PH Registry, you have
to show that *ALL* of the EDU.PH DNS servers were down for *some*
particularly significant time interal.
Mind you, your case is easy. You don’t have to show that *all* were down
*all* the time, just that *all* were down for *some* time. You even get to
choose the time interval you want to study and there are infinite numbers of
such intervals over a given 24hour period.
Next, let us assume that you are indeed able to find such an interval and
therefore prove that EDU.PH did have an outage. All you would have proven is
that EDU.PH did have an outage and nothing more. You wouldn’t have proven
that
PHNET can not handle the PHccTLD Registry. You wouldn’t have proven that
another organization can run the PHccTLD Registry. You would not have proven
that dotPH has a technically superior system
There are currently four DNS servers that the PHccTLD currently uses. They
are not exclusively for the use of Joel Disini nor exclusively for dotPH nor
for whatever company the Disinis wish to front for them. In fact, except for
UUNET, all of the servers provide DNS service for the PHccTLD because it is
the PHccTLD. In other words, at least 3 of these 4 servers will continue to
provide the DNS service to the PHccTLD — regardless of who runs the
registry.
Therefore, even if the Disinis were to prove that PHNET is not a good
registry for EDU.PH because of outages, the following do not follow:
(1) that PHNET can not be a good registry for the PHccTLD.
— The servers currently used by dotPH for the PHccTLD would also be
available to PHNET if it were to be the Registry. The servers would be
available to anyone who becomes the Registry.
(2) that dotPH is technically superior to anyone else in running the
Registry DNS
— The PHccTLD servers the Disinis rely on are not run by dotPH and
therefore, their uptimes and claimed superiority are not reflective of
dotPH’s technical capability.
(3) that no one else in the entire Philippines can take over the Registry
function from the Disinis
Sorry, Joel, you lose again. You can now join your lying brother in the
Posted by bombim cadiz at March 4, 2005, 4:24 pmloser’s corner and dialog with him.