Afrihost - Pure Fibre Feedback Thread

Status
Not open for further replies.

I use to see 10.189.30.18 now its 10.189.30.13. Think it has something to do with:

UPDATE 13:11, 22 JAN 2020
We’ve procured additional international capacity, which is being added progressively during the course of the day. We expect that this will greatly improve our client’s experience, and we will continue to monitor performance closely.
 

on its way.. ETA 28 Jan 11PM.
 
Has this issue ever been resolved? Still barely watchable here

If single thread is kuk, then RD will be kuk.

VPN:
8973321695.png


NO VPN:
8973328173.png
 
Hey guys

Sincere apologies - I was at school yesterday. I actually forgot about it completely and was on my way to the office. Lukily someone reminded me. I will answer all your posts and PM's in the next hour or so - I see there were a lot.

Really sorry about leaving everyone hanging. Promise I will get to everything.
 
It has to do with your ISP, as Cloudflare uses anycast or all their services. I suspect because the western cables are down and everyone's going via East, some ISPs are routing via JHB to the east coast. I know before the cable incident Afrihost did use Cloudflare's CPT location.

Maybe @AfriMan can assist here.

Both west coast cables are down, so pretty much everything is routing off the east coast. I'd have to get a little more info for specific services, like a traceroute, etc, so I can check with NOC.
 
Day #2 - @AfriMan or any other AH rep still on here? Any feedback on this?

Got some feedback from our Network Manager.

From our tests, Cloudflare is working properly. Here is a test from a CPT location to MyBB (hosted on Cloudlfare):

cpt ~]$ traceroute -q 1 -w 1 mybroadband.co.za
traceroute to mybroadband.co.za (104.20.10.169), 30 hops max, 60 byte packets
1 cpt-os2.ip.adsl.co.za (169.1.2.249) 0.754 ms
2 169-1-21-83.ip.afrihost.co.za (169.1.21.83) 0.086 ms
3 cloudflare.ixp.capetown (196.10.140.198) 0.128 ms
4 104.20.10.169 (104.20.10.169) 0.100 ms

So this means that any differences could be due to how Cloudflare sees us, but we peer at NAP Africa, so this should not be an issue. It's more likely that Cloudflare may choose to make some services only available in JHB, and routes all traffic there. Some IP's we tested were only available in JHB, so that is the most likely reason.
 
Got some feedback from our Network Manager.

From our tests, Cloudflare is working properly. Here is a test from a CPT location to MyBB (hosted on Cloudlfare):

cpt ~]$ traceroute -q 1 -w 1 mybroadband.co.za
traceroute to mybroadband.co.za (104.20.10.169), 30 hops max, 60 byte packets
1 cpt-os2.ip.adsl.co.za (169.1.2.249) 0.754 ms
2 169-1-21-83.ip.afrihost.co.za (169.1.21.83) 0.086 ms
3 cloudflare.ixp.capetown (196.10.140.198) 0.128 ms
4 104.20.10.169 (104.20.10.169) 0.100 ms

So this means that any differences could be due to how Cloudflare sees us, but we peer at NAP Africa, so this should not be an issue. It's more likely that Cloudflare may choose to make some services only available in JHB, and routes all traffic there. Some IP's we tested were only available in JHB, so that is the most likely reason.

Aah, that sucks, but thanks for checking.
 
Always happy to check. We'll also look from our side. If there is anything we think could be working better, we'll look into it.

No worries - guess it means Cloudflare only have a Warp endpoint in JHB for the whole of SA. Seems sub-optimal, but maybe they're still implementing.
 
I don't get the thought process then, unless whatever backbone/routing they're using goes through JHB instead of CPT, so having a CPT endpoint doesn't make sense.

I think it really depends on how Cloudflare does their routing and where their PoP's are.
 
I think it really depends on how Cloudflare does their routing and where their PoP's are.

Cloudflare's Warp via Web Squad appears to be working in Cape Town (via their LG):

[capetown.lg.as328137.net] trace to 104.16.85.15
# ADDRESS LOSS SENT LAST AVG BEST WORST
1 core.as-01-xe.cp1.za.ws.net.za 0% 1 0.1ms 0.1 0.1 0.1
2 core.cr-01.cp1.za.ws.net.za 0% 1 0.1ms 0.1 0.1 0.1
3 160.119.232.20 0% 1 0.1ms 0.1 0.1 0.1
4 cloudflare.ixp.capetown 0% 1 0.2ms 0.2 0.2 0.2
5 104.16.85.15 0% 1 0.2ms 0.2 0.2 0.2
 
Doing some remote work and decided to see how Afrihost compares to Warp vs Warp+

Afrihost:
8975670385.png


Warp:
8975681297.png


Warp+:
8975667378.png


Tests are to Telstra in Perth, Aus
 
Cloudflare's Warp via Web Squad appears to be working in Cape Town (via their LG):

We should be same:

cpt ~]$ traceroute -q 1 -w 1 104.16.85.15
traceroute to 104.16.85.15 (104.16.85.15), 30 hops max, 60 byte packets
1 cpt-os2.ip.adsl.co.za (169.1.2.249) 0.973 ms
2 169-1-21-83.ip.afrihost.co.za (169.1.21.83) 0.104 ms
3 cloudflare.ixp.capetown (196.10.140.198) 0.106 ms
4 104.16.85.15 (104.16.85.15) 0.088 ms
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X