All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
The IPv6 outage thread - New Year's edition 🎇
🎇 Happy New Year, everyone! 🎆
I wish each of you all the best, healthiness and broader IPv6 adoption in 2022.
But first, let's have a look back at how IPv6 performance has been in December '21:
✅ mrvm/WebHorizon – There have been a handful of issues with IPv6 connectivity on some of my NAT boxes, but all of them were handled quickly by the respective providers (or upstreams), so I won't go into detail here.
💩 HostMaze – While they didn't have any IPv6 outage last month, things became way worse this month. The first outage was on Dec 09 and since then there had been hardly a day without any issues:
(my email notification is triggering after 30mins of downtime)
Sometimes I got "no route to host"-error messages from their router, sometimes packets went in circles until their TTL expired and sometimes traffic just disappeared in the void. Currently, IPv6 is down since Dec 29. They've been surprisingly responsive at first and got IPv6 back up and running every single time, but I haven't heard from them since Dec 19. They've never given me any details about the issue so I don't know what's going on.
Even worse, they've migrated from SolusVM to Virtualizor in December (without advance notice) but my VPS doesn't show up in my account so I'm currently unable to manage it. My ticket about that issue hasn't been responded to yet.
👍 GSL Networks/ZappieHost – I stumbled upon an IPv6 routing issue in GSL's network on Dec 18 when I was unable to connect to my ZappieHost VPS from my residential connection at home (Vodafone). Other networks didn't seem to be affected. I opened a ticket with ZappieHost, who use GSL as their primary upstream, two days later. Adam responded promptly and even though this issue seemed to affect Vodafone Germany only, he spent some time with me to debug the issue and filed a ticket with his upstream eventually. The issue was resolved shortly after.
😞 GreenCloudVPS – My VPS with GreenCloudVPS in Vietnam started showing high packet packet loss on Dec 24. It took me a while to figure out what was going on since "normal" ICMP traceroutes didn't show any issues but mtrs using UDP or TCP showed parts of the outgoing traffic getting stuck in loops between two hops in their upstream network. I put all information I got together and opened a ticket with them.
They obviously didn't even bother to review the information in the ticket as I got two unrelated, general responses first:
During this time, Vietnam's international network is slower than other locations because of a cable break
In this case,you can check again in a few weeks
and
The network port/bandwidth is fair-shared. So sometimes you might get slower or packet loss. In this case, Please check again in few days
After both responses I updated the ticket and tried to make them understand that my issue isn't caused by congestion but by a misconfiguration in their upstream's network. Then they responded with:
Could you please give us the VPS's password for further check?
I swallowed my anger and created an account for them. They started fiddling around with ethtool and iptables, took a tcpdumpof the traffic coming from 8.8.8.8 and tried to run fancy stuff like:
sudo ipv6 icmpv6 unreachable
Then they stated:
We have checked and it seem like the network is still normal
I got somewhat upset and sent them a wall of text (which I apologized for later) telling them that I'm not under the impression that they've reviewed any of the information I provided them with and asked them whether whey wanted to continue working with me to get this resolved or not.
They responded dryly:
Please wait, we will check again
As expected, their "checks" didn't improve the situation at all, so I started another attempt to show them the impact of the misconfiguration, this time using iperf3 which provides great insights such as packet loss, packets arriving out of order and transfer speed, of course. But I was tilting at windmills:
In this case, please wait few days to fix the network cable
Or you can send us the MTR link, we can request the datacenter routing for you but we could not guarantee it can fix your problem
Well, I already provided them with MTRs, but I sent them some more. After that, they asked for traceroutes and after they got those, they asked for ping outputs. Then I was told...
Please wait, we will check and back to you soon
...and they closed the ticket. When I asked for an update two days later, they closed the ticket again and responded:
Our datacenter responded that it's working normally and sometimes the packet loss is just 1-2%. Unfortunately, it's our best route we can offer in Vietnam now.
I've called it a day and didn't push the ticket again as they obviously are either unable to understand the cause of the issues I am seeing or unwilling to get this resolved for whatever reason. I've been a happy customer of GreenCloudVPS so far, but this time I feel like they've been wasting my time right from when I opened the ticket.
👍 Central.so – Does not qualify as an "outage", but I asked them about their single-homed HE network in their recent offer thread and they've added a second upstream and connected to an IX few days later.
Comments
Someone please tell the OP, you cannot send packages over computer networks.
In German(y) we can.
Sorry, fixed. 🥺
Not with that attitude. Here are my packages sent to you over a computer network.
Or...
Or...
*apt/dnf/pacman wants to know your location*
Your IPv9 is lacking.
Thank you for the kind words
I can confirm that Vodafone have many routing issues in the last months.
I am also using a Vodafone cable connection at home.
In the last two months there are so many problems with routing , as I had the last two years with my old provider Telefonica.
That's verry bad. Vodafone is using the "aorta" international.
As I regnoiced the issues the packet loss began as I reached the aorta nodes.
The issues are only one or two days long..after that time the routing is running correctly.
The worst thing is that you're unable to talk to a sane person at Vodafone about such issues. Their support hotline will do some measurements and state that there's no issue with your line and if you insist you're having issues they'll replace your wifi router.
Quick update about my open issues: Neither Hostmaze nor GreenCloudVPS have responded to my tickets but in both cases IPv6 seems to be stable again.
Vodafone is dogshit, end of story.
The people behind the phone have a 4 to 8 weeks training, they don't know shit either.
Try twitter or other sorts of communication, attach mtr's.... and surely one will help.
If not, you just stop paying them, works like magic.
Yeah, tried that with DTAG and got my line suspended.
IPv6 is a dead cause. I am going to wait for IPv12.
First talk, then shoot.