Cool Ideas Fibre ISP - Feedback

[John Tempus]

Could you switch back to the dynamic again please ?

Im back on it, couldnt load any traffic over the static ip.

The best outcome I had before switching back is the following however without being able to load a single page.

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.0.1 - 0 | 49 | 49 | 0 | 0 | 0 | 0 |
| - 0 | 49 | 49 | 2 | 11 | 75 | 3 |
| 154.0.1.125 - 39 | 18 | 11 | 5 | 12 | 25 | 9 |
| 154.0.1.13 - 42 | 17 | 10 | 0 | 17 | 44 | 7 |
| 196.10.140.198 - 50 | 16 | 8 | 5 | 18 | 53 | 11 |
| 1.1.1.1 - 37 | 19 | 12 | 4 | 14 | 35 | 4 |
|________________________________________________|______|______|______|______|______|______|


vs dynamic ip

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.0.1 - 0 | 36 | 36 | 0 | 0 | 1 | 0 |
| - 0 | 36 | 36 | 2 | 10 | 40 | 2 |
| 154.0.1.125 - 0 | 36 | 36 | 1 | 1 | 1 | 1 |
| 154.0.1.13 - 0 | 36 | 36 | 1 | 1 | 3 | 1 |
| 196.10.140.198 - 0 | 36 | 36 | 1 | 1 | 3 | 1 |
| 1.1.1.1 - 0 | 36 | 36 | 1 | 1 | 3 | 1 |
|________________________________________________|______|______|______|______|______|______|

 

[TheRoDent]

Could you switch back to the dynamic again please ?

Thanks.

OK, this helps me a lot. Again, evidence. I can use. It makes no sense.

Our handoff is a single BGP session, in theory, this should all be the same.

A static assigned massive packet loss, and latency. A dynamic after hours, working well....

 

[John Tempus]

Thanks.

OK, this helps me a lot. Again, evidence. I can use. It makes no sense.

A static assigned massive packet loss, and latency. A dynamic after hours, working well....

Right, only difference is that static ip is from a different ip range. If all ips are treated equally then the outcome should be the same assuming they are assigned on the same switch.

The packetloss on that static ip looked to me like heavy shaping occurring for some bizarre reason.

 

[TheRoDent]

So, this is part of the network that we don't have insight to. The routing, and handoff for that IP should be exactly the same.

I get an interface 154.0.1.125 (my router) that I get fed packets from the Trenched network. What happens inside it is a black box, and I have a bit of a web interface to manage your MAC addresses, and static IP's.

The rest is entirely opaque to us as an ISP.

 

[TheRoDent]

Queue bun-fight with Vumatel. I had my daily dose with Octotel this morning.

 

[John Tempus]

So, this is part of the network that we don't have insight to. The routing, and handoff for that IP should be exactly the same.

I get an interface 154.0.1.125 (my router) that I get fed packets from the Trenched network. What happens inside it is a black box, and I have a bit of a web interface to manage your MAC addresses, and static IP's.

The rest is entirely opaque to us as an ISP.

The results is closer to the Twilight zone. o_o

 

[John Tempus]

Queue bun-fight with Vumatel. I had my daily dose with Octotel this morning.

Just a thought, is it possible they might be offloading certain ip assigned to different physical hardware ie. the static ip you assigned might have been offloaded on a faulty switch vs the dynamic ip being assigned on a lets-call-it functional switch.

Is this at all plausible ?

 

[TheRoDent]

The results is closer to the Twilight zone. o_o

Would you mind keeping ICMP open?

 

[TheRoDent]

Just a thought, is it possible they might be offloading certain ip assigned to different physical hardware ie. the static ip you assigned might have been offloaded on a faulty switch vs the dynamic ip being assigned on a lets-call-it functional switch.

Is this at all plausible ?

This is entirely plausible. But in theory the have a "managed" layer 2 network to our handoff in Bree street, but since they take care of DHCP, and routing, yes. It could be that the "static" network takes a different path.

They do the layer 3 on our behalf. I wish they didn't. Just to put it in perspective of how much of a "black box" it is. We have zero insight aside from the first (third) handoff hop.

With other FNO's (Telkom excluded) we at least get raw ethernet packets....

 

[John Tempus]

This is entirely plausible. But in theory the have a "managed" layer 2 network to our handoff in Bree street, but since they take care of DHCP, and routing, yes. It could be that the "static" network takes a different path.

They do the layer 3 on our behalf. I wish they didn't. Just to put it in perspective of how a "black box" it is. We have zero insight aside from the first hop.

PS. this is why at no point in this thread and the issues with Vumatel did I ever once blame CISP just so that we are all on the same page. My fight is with Vumatel and I appreciate the interaction coming from you guys.

 

[Phinix]

Has been rather fascinating reading through this all.
Suprised you aren't passed out in your bed by the sounds of what you've been fighting through @TheRoDent

 

[John Tempus]

Has been rather fascinating reading through this all.
Suprised you aren't passed out in your bed by the sounds of what you've been fighting through @TheRoDent

Going by his nickname I am pretty sure nocturnal hours is the most productive.

 

[TheRoDent]

Going by his nickname I am pretty sure nocturnal hours is the most productive.

Hah, I live on Heineken and good company. Sleep is a by product of hard work.

 

[John Tempus]

Hah, I live on Heineken and good company. Sleep is a by product of hard work.

Here is what I will do today after loadshedding when I am 99.99% certain the link will just go into blinking hell until probably the same time of the day. Will that static ip be usable for me to assign right after load shedding to see for myself even if the traffic is still wonky if the sync will occur and the ip assigned just enough to do the same MTR test.

I suspect going dynamic or static ip the line sync will simply not occur after loadshedding. I might be proven wrong or Vumatel might have fixed something in the wee hours just now, tomorrow will tell I guess.

 

[TheRoDent]

Here is what I will do today after loadshedding when I am 99.99% certain the link will just go into blinking hell until probably the same time of the day. Will that static ip be usable for me to assign right after load shedding to see for myself even if the traffic is still wonky if the sync will occur and the ip assigned just enough to do the same MTR test.

I suspect going dynamic or static ip the line sync will simply not occur after loadshedding. I might be proven wrong or Vumatel might have fixed something in the wee hours just now, tomorrow will tell I guess.

The static is assigned to your circuit, and doesn't require any DHCP involvement (in theory).

You can re-use it at any time. I make no promises however, as to it's veracity or usefulness but at least it's something we can eliminate.

If it works, or eliminates your down-time, then I will assign a static to every single one of our clients. It costs's me nothing. *edit* aside from some hard work, which I'll code away.

 

[John Tempus]

The static is assigned to your circuit, and doesn't require any DHCP involvement (in theory).

You can re-use it at any time. I make no promises however, as to it's veracity or usefulness but at least it's something we can eliminate.

If it works, or eliminates your down-time, then I will assign a static to every single one of our clients. It costs's me nothing. *edit* aside from some hard work, which I'll code away.

Yup that is the only thing I want to test during the blinking hellzone. If I assign the static would it cure the blinking hell and if so we need to figure out how to actually use the static ip since the current static ip test resulted in unusable connectivity.

 

[kripstoe]

I wouldn't consider this thread and the various notices we put up as radio silence. Do understand, if you were with another ISP, they would give you the same information.

I concede that there are some generic notices in the website, but so does every other ISP. But as an ISP of the year, there should be more visible/regular comms or at least the perception that more is being done to address or push Vumatel/Octottel/whoever on status updates. Why else would I use CISP, if it's the same as ISP XYZ?

Edit: Thanks for the update though

 

[kripstoe]

Who did you have to chase back in the days of ADSL? Sure, you could engage with your ISP, but everyone knew at the end of the day that only Telkom could fix your ADSL.

It's no different today. Just the names have changed from Telkom to Vumatel/Octotel/Frogfoot/<insert name here>

True, but ADSL is not Fibre. We would have hoped the tech and how me manage it would have improved with all the money being poured into getting fibre everywhere. In this specific case, it seems (perception) that it is worse than ADSL.

 

[TranQ]

Queue bun-fight with Vumatel. I had my daily dose with Octotel this morning.

Any update with regards to Netstream? Is there any plan from their side to fix whatever is wrong with my line? It's now been going on for 4 months and it is even worse than when I first reported it.

 

[Praeses]

I see Frogfoot decided to do unannounced scheduled maintenance from 5 this morning until 17:00 in my area. Hopefully it'll fix my twitch streaming