HE Fremont IPv6 tunnel

Mailman List Mirror (Read Only)
1

Recently I migrated the server that's running an HE IPv6 tunnel to one of the Fremont endpoints and now the tunnel is going down for a few minutes every couple of hours or so. I haven't been able yet to find a reason for this. I made sure the old server is not running any IPv6 related things anymore (such as radvd).

I am curious if anyone knows if the Fremont endpoints are experiencing some problems? I have barely ever noticed any problems with the IPv6 tunnels so I am pretty sure it's not on HE's side, but I wanted to make sure.

Thanks,
Jeroen

2

Perhaps you should try contacting HE support. I hear they're responsive.

3

Rob Mosher wrote:

Perhaps you should try contacting HE support. I hear they're responsive.

I understand however I was pretty sure it wasn't the tunnel that was the problem. So I didn't feel emailing HE was appropriate.

I am curious, since I have pretty much confirmed the problem is on my side, why would a move of an IPv6 tunnel from one server to another suddenly cause intermittent outages of a few minutes every couple of hours. It's not as regular so as to suspect a cronjob or something.

The old server is still online, however the radvd daemon isn't running there anymore and its external interface has been disabled. There is no IPv6 dhcp server running either.

The new server basically has the configuration and IP addresses the old one used to have. It's running radvd. Is there some (obvious) residual effect of having moved an IPv6 tunnel from one physical server to another that I fail to recognise?

I have moved ethernet interfaces and cables on the new server to rule out a fault in those. Its internal interface (on the same ethernet card) appears very stable.

Thanks,
Jeroen

4

Jeroen van Aart wrote:

I am curious, since I have pretty much confirmed the problem is on my side, why would a move of an IPv6 tunnel from one server to another

A helpful person pointed me in the right direction. Multiple times I checked the /etc/network/interfaces file and didn't spot it. But I had forgotten to remove the gateway for the internal interface... after switching IPs and enabling the external interface.

Stupid mistake, and obviously it sort of works, but not well.

Thanks,
Jeroen