MWEB on OpenServe packet loss

L

Leroy Brown

Well-Known Member
Joined
Feb 5, 2014
Messages
128
Reaction score
0
Location
Centurion
Been having a lot of trouble with my fibre link the last couple of days.
In the evenings between around 7 and 10PM, my link becomes completely unstable and useless. Traffic loss sits at around 20%+ and even though speedtests will show my full speed of 40/10mbps, international pages will take 2min+ to open.

Anyone else been struggling with this? I've seen one more person with the exact same issue on Carbonite.
We both get our packet loss at 197-80-96-161.jhb.mweb.co.za which is the 3rd hop for my outbound route.

@MWEBHelp pls

1586719188561.png
 

Attachments

  • 1586719120467.png
    1586719120467.png
    106.9 KB · Views: 10
Seconded,
have logged two tickets since the 10th,
called numerous times today and yesterday and the day before,
lead technician is always going to ''call me back''
 
StarWez said:
Seconded,
have logged two tickets since the 10th,
called numerous times today and yesterday and the day before,
lead technician is always going to ''call me back''
Click to expand...

Any chance you're also in the PTA East area and also getting packet loss at the same node (197-80-96-161.jhb.mweb.co.za)?
 
Leroy Brown said:
Any chance you're also in the PTA East area and also getting packet loss at the same node (197-80-96-161.jhb.mweb.co.za)?
Click to expand...
I am in centurion,
and yes **** that node.


Just finished testing with a vox account on same router and that is working fine



getting loss per below hops

| 197-80-96-161.jhb.mweb.co.za - 16 | 13 | 11 | 0 | 3 | 5 | 3 |
| te0-0-0-3.vic-p-1.optinet.net - 16 | 13 | 11 | 13 | 62 | 226 | 13 |
| 197-80-7-66.jhb.mweb.co.za - 16 | 13 | 11 | 13 | 48 | 98 | 13 |
| te0-3-0-2.trv-p-1.optinet.net - 6 | 17 | 16 | 13 | 43 | 97 | 13 |
| 197-82-7-2.dbn.mweb.co.za - 6 | 17 | 16 | 13 | 45 | 147 | 13 |
| 197-82-7-5.dbn.mweb.co.za - 6 | 17 | 16 | 13 | 45 | 148 | 13 |
| 196-10-141-198.ixp.durban - 6 | 17 | 16 | 13 | 45 | 147 | 14 |
|
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
 
Yep


this is 100% an mweb issue, and the fact that i get told everyday, after calling twice a day to follow up that the dude is going to call me back shortly hacks me off.

i have tested now using vox and mtn ftth.
working 100%
 
Can't see where that hop is I'm assuming hop 2 or 3 right? That looks like IPC
 
@cavedog this is my route. Even though it's saying I'm getting 20,8% loss on the 3rd hop, I'm now not dropping any more packets with a normal ping to 8888. It comes right as soon as it gets late.

1586725523622.png

Ping statistics for 8.8.8.8:
Packets: Sent = 111, Received = 111, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 626ms, Average = 17ms
 
cavedog said:
What were you doing a trace to that is peering in Durban instead of JHB?
Click to expand...
Thats WinMTR , great app for packetloss.

google.com, google.co.za, carbonite.co.za, amazon.com
tested all of those, all have packetloss
 
Leroy Brown said:
@cavedog this is my route. Even though it's saying I'm getting 20,8% loss on the 3rd hop, I'm now not dropping any more packets with a normal ping to 8888. It comes right as soon as it gets late.

View attachment 820387

Ping statistics for 8.8.8.8:
Packets: Sent = 111, Received = 111, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 626ms, Average = 17ms
Click to expand...

The hop you highlighted is Mweb IPC. That could be congested during peak time.

The loss you see there is not line or ISP loss but rater the router that did not respond to icmp. Loss would show right through the trace route.

Depending on how bad the ipc congestion gets loss is probably very severe and would not have suddenly happened. It's usually a systematic increase as the ISP sells more services or if it started soon after lockdown when traffic increased.

Good luck getting it sorted. It needs to be increased and Mweb or IS is going to need to pay quite a bit extra. The new IP Connect model is not yet in effect.
 
StarWez said:
Thats WinMTR , great app for packetloss.

google.com, google.co.za, carbonite.co.za, amazon.com
tested all of those, all have packetloss
Click to expand...

All those should not be peering in Durban if you are in Gauteng as your MTR suggests.
 
cavedog said:
The hop you highlighted is Mweb IPC. That could be congested during peak time.

The loss you see there is not line or ISP loss but rater the router that did not respond to icmp. Loss would show right through the trace route.

Depending on how bad the ipc congestion gets loss is probably very severe and would not have suddenly happened. It's usually a systematic increase as the ISP sells more services or if it started soon after lockdown when traffic increased.

Good luck getting it sorted. It needs to be increased and Mweb or IS is going to need to pay quite a bit extra. The new IP Connect model is not yet in effect.
Click to expand...

Its weird though, as I've only encountered this on certain links. My mates who are running over Mweb and Openserve arent all having the same issue. Thanks for your input
 
Leroy Brown said:
Its weird though, as I've only encountered this on certain links. My mates who are running over Mweb and Openserve arent all having the same issue. Thanks for your input
Click to expand...
They might have more than one IPC - MWEB had one and IS had one so don't know if they still like that.
If you can diagnose your PPPoE session then check the PPPoE server you connect to!
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X