Cool Ideas Fibre ISP – Feedback Thread 4

DYreX146

Well-Known Member
Joined
Nov 14, 2018
Messages
468
Higher than normal in KZN as well on your first hop? or are you just referring to KZN routing via JHB first? As they are very different problems.

I was simply referring to @S.Claus and a few other customers specific problem with Openserve adding a bunch of latency depending on the path.
It could possibly be using JHB as a return path for some destinations, I'm not 100% sure. All I know is there's an increase of about 5-10ms on certain servers.

Example: hops 3 and 9 on this trace, the rest are fine.

1621237044212.png

And literally every single hop on this one has extra latency:

1621237004463.png

Is this something you have mentioned to Openserve yet? Or is it related to something else?
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
It could possibly be using JHB as a return path for some destinations, I'm not 100% sure. All I know is there's an increase of about 5-10ms on certain servers.

Example: hops 3 and 9 on this trace, the rest are fine.

View attachment 1069405

And literally every single hop on this one has extra latency:

View attachment 1069403

Is this something you have mentioned to Openserve yet? Or is it related to something else?
Yes so return latency doesnt matter unless it affects the destination latency, if your destination is normal that's the effective latency between yourself and the host in question.
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
I'm on a 100/100MB package and after experiencing loadshedding this morning my fibre connection was restored, however speedtests show that my upload speed is only ~50MB instead of 100MB?

I tested with multiple Speedtest.net ISP's (local and international - tested with a wired connection), as well as the MyBroadband mobile app on my phone, and the results were all similar.

FNO_ Frogfoot, Gordon's Bay, Cape Town.

Any ideas? @PBCool

#COOL-20210517-684073

Thanks!!
Try restart your router and test again.
 

DYreX146

Well-Known Member
Joined
Nov 14, 2018
Messages
468
Yes so return latency doesnt matter unless it affects the destination latency, if your destination is normal that's the effective latency between yourself and the host in question.
Well as you can see the destination latency of the 2nd trace is 4ms higher than the first one, so it is getting affected in some way.

But is this something you have taken up with Openserve yet?
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
Well as you can see the destination latency of the 2nd trace is 4ms higher than the first one, so it is getting affected in some way.

But is this something you have taken up with Openserve yet?
If you look at @S.Claus' situation his latency has a variance of around ~12ms and Openserves feedback is that this is fine, in your case a 4ms variance Openserve wouldn't even entertain investigating.

Possibly once we have some traction on @S.Claus' situation we can raise these other ones, it's their underlying load balancing between metros that is causing this, IE some packets take the one route of their ring, other packets take another path which is much longer.
 

DYreX146

Well-Known Member
Joined
Nov 14, 2018
Messages
468
If you look at @S.Claus' situation his latency has a variance of around ~12ms and Openserves feedback is that this is fine, in your case a 4ms variance Openserve wouldn't even entertain investigating.

Possibly once we have some traction on @S.Claus' situation we can raise these other ones, it's their underlying load balancing between metros that is causing this, IE some packets take the one route of their ring, other packets take another path which is much longer.
He noticed his issue not long after I pointed mine out, I just hope they are related in some way so that Openserve fixes them both at the same time.

But today's example is pretty mild compared to when I previously showed that I had 24ms to Durban as opposed to the 3ms I usually get, which is 21ms variance.

It seems to change day by day, right now I have a variance of about 7ms to DBN and JHB depending on the server as you can see here:

Discord (JHB) - 19ms
1621243016480.png

Google (JHB) - 12ms
1621243299740.png
 

TheRoDent

Cool Ideas Rep
Joined
Aug 6, 2003
Messages
5,575
Been battling with CoolIdeas since yesterday. Now determined that every time I get IP address 102.132.205.3 or 5, I show connected but no data flow. If I reboot the router enough times and get lucky, IP changes to 102.132.209.17 and then works awesome!! Problem is that after about 30 minutes or so, the router reverts to the 205.3 or 5 IP and I am dead again.

Any suggestions please?
@TheRoDent
@PBCool
Thanks, I got your PM. We're investigating this in cooperation with OpenServe.
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
He noticed his issue not long after I pointed mine out, I just hope they are related in some way so that Openserve fixes them both at the same time.

But today's example is pretty mild compared to when I previously showed that I had 24ms to Durban as opposed to the 3ms I usually get, which is 21ms variance.

It seems to change day by day, right now I have a variance of about 7ms to DBN and JHB depending on the server as you can see here:

Discord (JHB) - 19ms
View attachment 1069703

Google (JHB) - 12ms
View attachment 1069731
So one was a routing issue which we fixed, the other is layer2 related which is Openserve.
 

Armizael

Senior Member
Joined
May 31, 2006
Messages
526
Connectivity just dropped and when it came back up, latency is higher - ping to queen.cisp.co.za was normally ~150ms, about 200ms now:

Code:
Tracing route to queen.cisp.co.za [62.233.65.195]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  192.168.0.254
  2     *        *        *     Request timed out.
  3     3 ms     2 ms     2 ms  192.168.10.5
  4     1 ms     2 ms     1 ms  192.168.11.6
  5     2 ms     1 ms     1 ms  100.98.0.2
  6   198 ms   198 ms   199 ms  100.99.2.5
  7   200 ms   202 ms   201 ms  100.98.1.98
  8   199 ms   201 ms   199 ms  be20.asr01.thn.as20860.net [195.66.224.207]
  9   201 ms   201 ms   201 ms  130.180.204.33
10   204 ms   204 ms   203 ms  94.zone.4.r.dc9.redstation.co.uk [185.20.96.154]
11   202 ms   203 ms   201 ms  queen.cisp.co.za [62.233.65.195]

Looks like it happened at 18:08.

Openserve, CPT.

EDIT: Back to normal now.
 
Last edited:

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
Connectivity just dropped and when it came back up, latency is higher - ping to queen.cisp.co.za was normally ~150ms, about 200ms now:

Code:
Tracing route to queen.cisp.co.za [62.233.65.195]
over a maximum of 30 hops:

1  2 * * * Request timed out.
3 3 ms 2 ms 2 ms 192.168.10.5
4 1 ms 2 ms 1 ms 192.168.11.6
5 2 ms 1 ms 1 ms 100.98.0.2
6 198 ms 198 ms 199 ms 100.99.2.5
7 200 ms 202 ms 201 ms 100.98.1.98
8 199 ms 201 ms 199 ms be20.asr01.thn.as20860.net [195.66.224.207]
9 201 ms 201 ms 201 ms 130.180.204.33
10 204 ms 204 ms 203 ms 94.zone.4.r.dc9.redstation.co.uk [185.20.96.154]
11 202 ms 203 ms 201 ms queen.cisp.co.za [62.233.65.195]

Looks like it happened at 18:08.

Openserve, CPT.

EDIT: Back to normal now.
Yes we have seen three flaps via WACs so those spikes are then routing via Eassy. We are awaiting feedback from carrier.
 

luckystrike

Well-Known Member
Joined
Nov 13, 2007
Messages
100
@PBCool @TheRoDent @2cool4u

Hello,

My internet down in Lynnwood Pretoria this morning on Openserve. Is there a problem that you know of?

I seem to have the same IP range that @2cool4u had problems with.

It's connected but I have No connectivity. I tried restarting the router a bunch of times but I just get one of these 3 ip's.

102.132.205.4
102.132.205.5
102.132.205.6

I can PM you my details if it's not a general problem in the area?

Kind regards

Update:
Restarted my modem again and i got an ip in the :102.132.232.... range and internets back up and running.
 
Last edited:

CrypticZA

Expert Member
Joined
Sep 21, 2019
Messages
1,721
Any news from Amazon? VPN is great and all but it creates a bunch of additional issues which you need to work around which are rather annoying to deal with, and if we don't use it i die sliding behind walls.
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
Any news from Amazon? VPN is great and all but it creates a bunch of additional issues which you need to work around which are rather annoying to deal with, and if we don't use it i die sliding behind walls.
No further updates as yet, but based on the recent feedback here it seems as if they are still busy with changes. Where are things at the moment?
 

CrypticZA

Expert Member
Joined
Sep 21, 2019
Messages
1,721
No further updates as yet, but based on the recent feedback here it seems as if they are still busy with changes. Where are things at the moment?
We had one day where the route was correct other then that nothing has been done. Every night between like 8 and 9 ping to Frankfurt in game spikes to 210 so it makes experience even worse
 

PBCool

Cool Ideas
Company Rep
Joined
Jan 11, 2016
Messages
10,473
We had one day where the route was correct other then that nothing has been done. Every night between like 8 and 9 ping to Frankfurt in game spikes to 210 so it makes experience even worse
Cool I've updated AWS with your feedback, I will update with any responses.
 

luckystrike

Well-Known Member
Joined
Nov 13, 2007
Messages
100
And i have that 20 ms ping to ixp.joburg again. That wasnt there yesterday. It also makes my ping the same to bahrain as the capetown guys.

Attached yesterday and today.
 

Attachments

  • aws Frankfurt monday.png
    aws Frankfurt monday.png
    53.9 KB · Views: 17
  • Frankfurt18052021.png
    Frankfurt18052021.png
    32.8 KB · Views: 17

CrypticZA

Expert Member
Joined
Sep 21, 2019
Messages
1,721
Cool I've updated AWS with your feedback, I will update with any responses.
I honestly wouldn't be complaining if it was something small but it honestly just doesn't make me want to play the game, seeing some weird routing from Google to their Belgium server where the traceroute goes to 210 (normal and VPN is 180 going via London) and then i deal with AWS rouge routing as well so most nights its a gamble whether i get 140ms/150ms (London/Amsterdam) or 210 since both Google and Amazon EU routing has been questionable...

Not sure what is up with Google but its 140 to London (via VPN) and then another 30-40ms to Belgium on some servers.

And well Bahrain i can sit in queue with the 8 other people on the server.
 

MDE

Expert Member
Joined
May 18, 2009
Messages
2,050
Does anyone have any problems accessing or navigating ABSA online banking. Quite often get errors that it took to long to access the API. I know that they host on AWS, so just wondering if it's not related.
 
Top