What happened Monday night?

[MrGray]

Could only access the Sentech website for much of Monday evening - Sentech have put a notice on their fault notification board blaming their upstream provider's international link (IS presumably). I'm a little sceptical about that though as local was very patchy and for at least an hour all I could get was Sentech's site, and then not even that for a while.

If international circuit really was down why was half the local net either inaccessible or so slow as to be unusable? Could reach IS hosted local sites only for a while, and nothing anywhere else local - anyone know what was going on?

 

[Brolloks]

Don't know what happened, but I downloaded a 700MB file from IS at 13.2KB average. Best I've had ever. Used Getright with 5 consecutive threads. Total time 15h05m.

MyWireless 128, Tower: 82, Signal: 13%, SNL: 6, ber: 42%, node: 3

 

[nonroker]

Yeah..I actually got 27KB/s from ftp.is.co.za
This is the first time EVER

Ofcourse, now that intl is back up, it seems that my local speeds are screwed again

--
256k ... BAH..more like 25.6k
FSCK YOU Sentech!! [!]

 

[LethalChicken]

Hi Guys, not on Sentech but on Satellite, I was unable to use the International link as the IS connection to the UK was down .. so this might have affected you too ...

We are Telkom - Resistance is Futile - You will be Assimilated

 

[donn]

What I don't understand is this: the internet is *supposed* to be fault tolerant: if one route breaks then another route will kick in. So why doesn't it work in practice?

Surely people like IS have a system in place (at least for DNS) whereby if they can't link via their precious undersea cable to Europe, then can still get DNS information from, say, Australia, or the US. I wouldn't expect this kind of sophistication from Sentech, Tiscali or DataPro, but I would expect it from "the big boys" like IS, or is this just wishful thinking?

<blockquote id="quote"><font size="1" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"> From http://www.datapro.co.za/customerzone/
What is happening on the DataPro network?
IMPACT: All JHB customers

START TIME: 15 September 2004 15H20

SERVICE EFFECT: International connectivity

UPDATES 12H10: An upstream service provider is experiencing a problem with their International link between Rosebank and London. Their engineers are working to resolve it.
<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">

You can tell who the amateurs are: Sentech had nothing on their pages when the problem started (was I too quick off the mark maybe?), and DataPro still have a message from last week.

Does IS have a public page that lists their network status? Or do we have to wait for the <b>Sleepy Sentech Support</b> to tell us what is going on?

<hr noshade size="1">
Donn Edwards
<div align="right">Just because they <b>say</b> it's broadband doesn't make it so</div id="right">

 

[LethalChicken]

I think the general attitude is the same .. "You don't like our service, HAHAHA .. where else are you going to go, so shut up and bend over" ...



We are Telkom - Resistance is Futile - You will be Assimilated

 

[Dean_Henstock]

IS had a 45mb circut out Last night.

The internet is fault tolerent as long as you use two peering links, however Sentech only use 1 for International (being IS) therfore when there link went down so did all international DNS for Sentech.

<font color="red"><font size="1">If you fill up a station wagon with data tapes and drive to capetown you could transfer 10000 GB in 11 hours, and it would cost you R 500 in petrol, Thats the kind of internet I wanna see !</font id="size1"></font id="red">

 

[MrGray]

I know IS had an international circuit outage, but what I don't understand is why much of the local web, i.e. iol.co.za, news24.co.za, etc, were also not working for some time, and even sentech.co.za briefly.

There are more things in heaven and earth than are dreamed of by mere mortal men - Shakespeare

 

[loosecannon]

because sentech are amatures and dont know how to ensure there system is resiliant ...

DNS admin 101 ensure redundancy FFS use a dialup to mweb to feed your DNS servers never cut them off you will loose all sites this way and the problem will look worse than it is ...

the fact that there own site was affected is criminal ...

 

[Dean_Henstock]

LOL, now thats the funneist **** I have hear, tru but funny, they should colocate a mywi at a ISP and vpn to it, that will give them some redundancy and only cost R 600

lol


<font color="red"><font size="1">If you fill up a station wagon with data tapes and drive to capetown you could transfer 10000 GB in 11 hours, and it would cost you R 500 in petrol, Thats the kind of internet I wanna see !</font id="size1"></font id="red">

 

[dorris]

The problem was NOT limited to their DNS servers, bandwidth was the issue at hand, it just so happened that when the circuit died, I was doing a ping to google, and therefore had the IP in my terminal windows, copy pasting into firefox did nOT open google, a simple DNS Backup route would not have sufficed this fsckup, they need a backup routing table

 

[MrGray]

I can confirm it wasn't the DNS - I was quite happily resolving IP's for everything just no connectivity, even temporarily to Sentech's own site. It also seems that some local websites are still extremely slow while many are very fast. Anybody else get this?

 

[loosecannon]

the use satalite for the return path for via there proxy ... and the outgoing is via IS some local sites died because DNS timed out more than BW but there was a bandwidth failure that if they had some basic redundancy in place would have kept things contained to int only and not local ... they twits and blameing IS is good and well but what did they do to avert/minimise the impact on there customers in the intrest of good service.

 

[LordVader]

E-mail was up, during all this. I was having an e-mail conversation ,crapping them out... long story... with one of the dudes at the helldesk. I realized international was down, but didn't say anything. At about 10, I got an e-mail from the same dude at the helldesk, telling me that he just wanted to let me know that IS's link to New York was down, and he was waiting for a timeframe from IS for resolution on the problem. I believe everything was back up at around 1AM.

LordVader