back to

{{{ ANSW: First, one of the nodes in the line will attempt / succeed in establishing a connection to another node in the line 2 or more hops away. Maybe it works fine for a bit until someone picks up a 2.4GHz phone, or turns on the microwave... then the route goes dead. This issue popped up on me several months ago when the leaves fell out of the trees. I have nodes that would talk clear across town. How do you fix this? Easy: Log into each node and see what the signal strength is for each of the neighbors that you need that node to talk to. If you see links that show data being passed, but reporter shows a low value for the receive strength, then consider that to be a poor link. Set the minsig value to keep that node from talking to its neighbor nodes with poor signals. If there is some occasion where a marginal link has a high enough value to not meet minsig criteria, then put that node into the blocked nodes section of Wiana. I have found that if you use the blocked nodes section, it's best to put values in at both nodes. In other words, make Wiana apply the block from both ends. Other things to look at:

1. Make sure ALL nodes are running the SAME version of software. i.e., if your mesh is on Pro, upgrade all nodes to the latest version, v32.1925. If you are on OSS, upgrade everything to a good, stable version. I've not looked at anything in OSS beyond dev76 very much...

2. If this happens every hour, check to see how often the unblock retest is set for. Make it something like a day, or turn it off.

3. I turn off the "auto" power setting and set each node accordingly. If it has a 200mw card in it, I set it for 200mw. If it has a 30mw card in it, I set it for 100mw. Don't ask me why the 30mw cards get a 100mw setting - I know they won't put out that much, but they work better :) - I know number 3 isn't mutually agreed upon, but my mesh gets very wobbly when the power is set to auto, and remains stable at fixed levels.

QUES: > Does anyone ever have customers complaining about tiny 2 – 3 minute > outages? I have 1 segment of the network that suffers from this > problem. It is a 4 hop deep linear mesh. What happens is that every 1 > hour or so it loses all VPN inter-mesh tunnels. When I do a pingtest I can ping the 1.x address from the gateway but not the 172.16 VPN > tunnel IPs. By the time I realize the problem, ssh into the g/w node > and then into the problem node(s), the problem is gone.

I have host-mapped public IPs to all of my IPs and have a program that pings them all every minute. It pages/emails me if it fails 3 times in a row and then pages me again when it succeeds. From the 2nd hop node to the 4th, I get paged over and over again. Within a few minutes it goes out and then comes back. If I set my pager trigger to 5 missed pings, I get fewer failures but still a lot.

There is no way the wiana can report such tiny outages and I can only tell by our hardcore work-from-home Silicon Valley software engineers who complain very loudly when their coding get interrupted. I have been staying logged into nodes in attempt to see this happening. Mostly I can only see that my ssh widow is frozen, to come back and work in a few minutes. But today.. I got lucky and was in a node playing with tracroute and the node lost it VPN tunnel. It was the 3rd hop node and I logged in and jumped from 1 – 2 – 3 nodes using the direct neighbors 1.xx to ssh. So I didn’t loose my ssh, and when I did a traceroute to the g/w, The route took 1 hop to the 4th (not the 2nd) and the died. This next time I ran a traceroute again it went the proper way. So this is very strange, it doesn’t seem to happen on the entire mesh, but definitely happens here, and It makes me think that there is something causing the mesh box the dump it routing table and start over, or something }}}

Ad blocker interference detected!

Wikia is a free-to-use site that makes money from advertising. We have a modified experience for viewers using ad blockers

Wikia is not accessible if you’ve made further modifications. Remove the custom ad blocker rule(s) and the page will load as expected.