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 - Halloween edition 🎃
Besides a few nerds on this forum, no one gives a fuck about IPv6.
-- @DataRecovery, Oct. 2021
Motivated by the statement above and due to the great success of its predecessor, here's the second edition of the infamous IPv6 outage thread.
Few introductory words: One provider, which I won't call out, reached out to me after the first thread was published and didn't seem to be happy about being mentioned.
Dear providers, you will only be mentioned if I have a service with you and if you had IPv6-related issues in the past month. If you don't plan to maintain your IPv6 network properly, then stop offering service which have IPv6 enabled. Then I won't buy from you, you won't be mentioned and you'll save both of us the hassle.
Oh, and if you are clueless about what "gateways" and "default routes" are, then please switch to a different industry.
So let's go:
🤦♂️ Hostmaze -- They had their monthly outage on Oct 3rd this time. (Not an IPv6 outage in particular, but I'd like them to become a meme in these threads...)
🤷 Host Hong Kong -- I was wrong about them last month: They lose their IPv6 announcements more frequently. This time it got fixed in about 20 minutes.
✅ Owned Networks -- They had a full network outage in Lenoir on Oct 8. IPv4 came back online after about 24h iirc, but IPv6 connectivity was quite flaky. It got fixed within 48 hours after opening a ticket. They even offered me migration do a different location in case there was no quick fix for the issue.
✅ mrvm -- IPv6 connectivity was interrupted in Auckland, NZ due to an upstream issue, as they say. It was fixed in about 24h after opening a ticket.
🔻 AlexHost -- I've opened a ticket after noticing that I was unable to reach my Google Drive over IPv6. They didn't seem to really care about the information I gave them in the ticket and went straight to asking for my root password. ❌ First strike. I really, really dislike providers asking for my root credentials just to run two or three pings and traceroutes on their own. Wouldn't it be much more useful if they spun up a fresh VM on the same host to do their checks and see if they get the same results?
Surprisingly, they also couldn't ping Google from my VM. We went a bit back and forth and it turned out that they decided to run their IPv6 network with Cogent as their only upstream. ❌ Second strike. Running a production network with a single upstream is a bad idea already, but running a single-homed network upstreamed by Hurricane Electric or Cogent is even worse due to the ongoing peering disputes which are going on for at least 10 years.
They could have added a second upstream for IPv6, just like they have for IPv4, but instead they claimed: "we can't help you with that one." Considering the few bucks I spend with them, I cannot expect them to add another upstream just to make me happy, but getting this done would have resolved the issue for all customers and improved the quality and reliability of their network.
🤦♀️ Hostmaze -- Hostmaze? Again? Yes. Another outage on Oct 22. I'm amazed.
✅ FreeRangeCloud -- FRC had IPv6 hiccups in Ashburn once or twice a week and a longer outage on Oct 26 for unknown reasons. They were able to fix it by restarting their router and since then it looks stable.
🔻 HostForce / SolidSpace -- I have a VPS with them for years which has been idling along nicely so far. On Oct 26 it lost its IPv6 connectivity and outgoing traceroutes stopped after the first hop. I didn't get any response to my ticket until I pestered them in their Live Chat yesterday. "your ticket got stuck in the queue" - sure, that's what tickets do. They've asked for more traceroutes (spoiler alert: all of them stopped at the first hop) and then claimed they tried to replicate the issue from "each point in the network" and couldn't replicate the issue. So this must be my firewall then. I asked them to confirm the subnet assigned to me and the gateway IP and didn't get any response since then.
In summary I'd say that the month started rather slowly but became quite interesting in the end. There have been a few full outages as well, but I'm not mentioning them here since they're unrelated to IPv6. Let's see how the HostForce saga goes on. And I hope for some decent UltraVPS deal to replace my service with AlexHost.
Finally, here's the shout-out to all providers I got new services from which had IPv6 enabled and working right from the beginning: Zappie Host (they got my BGP session set up in 22 minutes -- world record 🏅), meerfarbig, WebHorizon, FriendHosting,
Happy Halloween!
Comments
view this post via IPv6 now.
Could be wrong but this one's quite easy to figure out:
1. No longer mentioned in this thread
2. Responding to ticket only after they got publicly called out
3. There were only 4 3 (-Hosterlabs) hosts which got meh/bad reviews in the last thread
Flat out refusal + service cancellation would be the proper response
Hostodo (I
Tostada) isn't listed?
I'm still seeing reachability issues early this month.
It might be my NDP setup but it's the same setup as elsewhere.
I lost hope and added SIT tunnels.
It's not that obvious, I'm afraid.
I thought so as well, but I couldn't find an email about a ticket, so I didn't list them. My monitoring shows an 1.5 hour outage on Oct 3rd, but it was during the night.
Hard to imagine anyone else on that previous list PMS-ing over getting called out, but I guess you really don't get to know your providers that well.