Excellent observations! But I would note that the TBB BQM graph for the Cerberus gateway that originally failed is now flat:

https://www.thinkbroadband.com/broadband/monitoring/...

Edited by shortshrift27 (Fri 06-Nov-20 10:43:43)

Yes IPv6 on the temporary gateway we've been swapped to is seemingly congested, flat green overnight rising up during the day and falling again. It's been a bit better since lock down which I guess is because some businesses have closed and staff are at home on more residential ISPs.

The IPv4 temporary gateway is as flat as a pancake though for me.

Are your BQMs to the temporary gateway we've been swapped to?

My IPv6 to gateway 46.37.56.2 (Gateway we were moved to)
https://www.thinkbroadband.com/broadband/monitoring/...

IPv4 to gateway 46.37.56.2 (Gateway we were moved to)
https://www.thinkbroadband.com/broadband/monitoring/...

Note the large spike at around 4:45pm on both of my charts was me dropping WAN then connecting again, just to see if I went back to the fixed gateway yet.

Before the latency issue on the original gateway IPv6 was quite flat.

For @lincsat their issue seems completely different to anything we've seen recently, certainly not related to the latency spikes.

Edited by E300 (Fri 06-Nov-20 10:52:15)

In reply to a post by shortshrift27:

Excellent observations! But I would note that the TBB BQM graph for the Cerberus gateway that originally failed is now flat:

https://www.thinkbroadband.com/broadband/monitoring/...

Yes definitely looking fixed now and you couldn't get much better, looks like the gateway and the Thinkbroadband monitor are in the same building

In reply to a post by E300:

Are your BQMs to the temporary gateway we've been swapped to?

I'm on 46.37.48.1

Text
1 23 45 67 89 1011 12	> /interface pppoe-client monitor 0           status: connected          uptime: 2d6h48m55s     active-links: 1        encoding:     service-name:         ac-name: acc-aln2.sv           ac-mac: 14:7B:AC:B2:F7:12             mtu: 1500              mru: 1500   local-address: x.x.x.x   remote-address: 46.37.48.1

(Session uptime because I was kicked off during the maintenance around 04:23-04:31am on 4th Nov)

In reply to a post by candlerb:

I'm on 46.37.48.1

Ah so on the fixed original one.

Odd you have those bottom to top spikes now and again, those don't show on the BQM against the gateway itself, perhaps something more local to you?

It's disappointing to see so much yellow on IPv6 on the original gateway during the day, I was hoping it was just the one we were switched to, perhaps a bit overloaded because it has more on there than normal. IPv6 didn't look like that before these issues on the original gateway, it was pretty flat, but now appears congested.

Although... is the IPv4 gateway address a good indicator of the IPv6 gateway, i.e. are they always in a pair or could they be completely different gateways? For example they have moved your IPv6 to a different gateway, but IPv4 has remained where it was? I only see a local IPv6 address for my gateway in pfSense, it did change at the swap over from what it was before, currently it is fe80::72e4:22ff:fed3:8c00.

In reply to a post by E300:

Although... is the IPv4 gateway address a good indicator of the IPv6 gateway, i.e. are they always in a pair or could they be completely different gateways? For example they have moved your IPv6 to a different gateway, but IPv4 has remained where it was?

It's a single PPPoE session, so both v4 and v6 must terminate on a single device.

I only see a local IPv6 address for my gateway in pfSense, it did change at the swap over from what it was before, currently it is fe80::72e4:22ff:fed3:8c00.

That's a link-local IPv6 address, derived from MAC address 70:e4:22:d3:8c:00 - and 70:e4:22 is a Cisco MAC address block.

Interestingly, the peer I see is 14:7B:AC which is Nokia.

I disabled IPv6 last Week when the issues started, I'm on the 46.37.56.2 gateway. It did settle down a lot in the early Hours

https://www.thinkbroadband.com/broadband/monitoring/...

Here is the snapshot from the 4th where this latest bad connection started

https://www.thinkbroadband.com/broadband/monitoring/...

Your issues do not look like anything we've seen during the problem Cerberus had and it's not what I'm seeing on the same gateway as you with IPv4 that is just green flat land.

The large amount of packet loss you were having yesterday and high latency seems to me to have been a fault local to you that then got fixed in the early hours. You still have a lot of yellow but better now as all the packet loss has gone so I expect the connection is working fine for you. The yellow can be caused by you just using your connection or by a router giving packets a low priority as well as congestion. Do you have any charts from a while back you can compare to in order to see if this is typical or not for you?

I would suggest waiting to see what happens when you move back to the fixed gateway (assuming we all do move back) and then if issues persist raise a ticket. It does seem likely though it is a problem more local to you, perhaps capacity issues somewhere that might suddenly go away when extra capacity gets added.

I assumed this latest issue was down to them moving everyone to the same gateway and now they are moving people away. Everything seems to be working OK.

Here is a graph from a Month ago showing how flat it used to be

https://www.thinkbroadband.com/broadband/monitoring/...

I'm on the same gateway as you and my IPv4 chart is just green, so for you to be seeing all that extra yellow means it is likely happening outside of Cerberus control. It might be that the fault you had yesterday meant somewhere some capacity has been lost until a permanent fix is made.

IPv4 on 46.37.56.2
https://www.thinkbroadband.com/broadband/monitoring/...