-
Notifications
You must be signed in to change notification settings - Fork 670
After running weave launch I only see a ipv6 listener #479
Comments
If you are getting that by doing something like
Please give more details of what you are doing when you try to setup a network. |
I have weave launch on two machines running.
And I do see the interfaces:
I am just trying to put together a little test cluster with 4 EC2 boxes using weave and I can't get two nodes to talk to each other. Unfortunately I am not quite sure where to see why the traffic is being dropped other than seeing if there is a tcp listener on 6783 (which is not there) |
Could you stop weave (do If that doesn't work, you probably need to change the EC2 security group settings to allow the machines to talk to each other on that port. Note that weave needs both TCP and UDP open on 6783. |
Is this related to #264? |
It seems to be related. I tried to reboot the boxes and it's still not working. Checked all of the sysctl values an they are set the way #264 discusses. It's like the packets are not being nat'd correctly into the weave network. What's weird is that there is a docker-proxy supposedly listening to 0.0.0.0:6783 but then the nat table also forwards the packets.
Also
But no traffic goes that way and it's really hard to figure out why not. |
Did the nc connectivity test work or not? |
You may also want to look at |
Yes that works. BTW: the command for listen needs -p node1:
node2:
|
To the other question...
|
and |
Sorry since it's says Reconnects, I didn't think you needed that. It is weird it says "ethwe" instead of "weave" for the network and what's the &{20?
|
That all looks fine. "no route to host" is the error seen by weave on the connection attempt. |
The other machine just in case you're interested:
|
Right, but I cannot get any further. No connection is being accepted by weave on the first machine due to some kind of routing problem. I can connect to port 6783 on localhost on the first machine just fine. I just cannot connect to port 6783 coming in from any other interface than local. |
on the second machine
just hangs. On the first machine
|
why does your weave bridge have an IP address, i.e?
More worryingly, the netmask is a /8. This overlaps with the eth0 interface. Did you run |
Ah, I didn't notice that. I was using an example and that might be the exact problem. Let me change that. |
That was it. Rookie mistake. I appologize. Awesome catch @rade by just looking at a few comments. |
Glad to hear that fixed it. I've raised #480 for making such mistakes easier to spot. |
btw, you said you were following an example. Which one? We really shouldn't have any examples that show exposing a /8. |
Another problem this example is now causing is that my interface is called "weave" but weave launch-dns is saying there is no interface called docker0. I guess this example is causing havoc with my brain :(. |
tcp6 0 0 :::6783 :::* LISTEN
It's not showing a tcp listener to 6783.
When trying to connect from another machine to setup a network I just get a timeout.
Not quite sure how to debug this to figure out what or where something went wrong.
I am using ubuntu 14.04
weave script 0.9.0
weave router 0.9.0
weave DNS 0.9.0
weave tools 0.9.0
Client version: 1.5.0
Client API version: 1.17
Go version (client): go1.4.1
Git commit (client): a8a31ef
OS/Arch (client): linux/amd64
Server version: 1.5.0
Server API version: 1.17
Go version (server): go1.4.1
Git commit (server): a8a31ef
Thanks
Andy
The text was updated successfully, but these errors were encountered: