back to http://scratchpad.wikia.com/wiki/Sasecurity
adsf[]
{{{
I need to clarify my response about "detectmode". From the sounds of things, your Wiana account was not properly configured? The dectectmode command will speed things up on detecting a gateway, but not if Wiana a is not properly setup. From the sounds of this, your gateway on the failed DSL router was still announcing itself as a gateway. > Has anyone found a neat way of switching internet gateways? > > I have a scenario today whereby an ADSL router has failed. I wanted to > switch to using the other gateway (satellite), which is hung off > another mesh node. > > I tried just rebooting the mesh node with the ADSL router, but all the > other nodes just waited until it came back up again! I tried > unplugging the router, but the mesh node still thought it was a > gateway. > > In the end I had to power off the mesh node and then all other nodes > found the alternate gateway. This still leaves me with the mesh node > attached the the ADSL which will not use the other gateway (cos it > thinks it still has one!). > > Is there neat way of forcing a switch? >I've added a wiki entry to explain how to utilize the feature to do this: >http://locustworld.com/tracker/wiki?p=DisableFailedUpstream >You may also want to avoid using the "Lock to gateway" option if you have that in place and utilize the "Preferred Gateway" setting if you're not using it already. >>Has anyone found a neat way of switching internet gateways? >>I have a scenario today whereby an ADSL router has failed. I wanted to switch to using the other gateway (satellite), which is hung off another mesh node. I tried just rebooting the mesh node with the ADSL router, but all the other nodes just waited until it came back up again! I tried unplugging the router, but the mesh node still thought it was a gateway. >>In the end I had to power off the mesh node and then all other nodes >>found the alternate gateway. This still leaves me with the mesh node >>attached the the ADSL which will not use the other gateway (cos it >>thinks it still has one!). >> >>Is there neat way of forcing a switch? I think the "detectmode" command will do this. However, I've once seen my mesh automatically pick up the backup gateway when our T1 went down. I wonder why yours does not? This happened to me when I was using the OSS - build25, dev76 I believe. It sounds as if you have a Wiana setting making that unit on the router think that it's always a gateway. Have you checked that, or does it have a static IP? Those things will make it think it's a gateway. > Has anyone found a neat way of switching internet gateways? > > I have a scenario today whereby an ADSL router has failed. I wanted to > switch to using the other gateway (satellite), which is hung off another > mesh node. > > I tried just rebooting the mesh node with the ADSL router, but all the > other nodes just waited until it came back up again! I tried unplugging > the router, but the mesh node still thought it was a gateway. > > In the end I had to power off the mesh node and then all other nodes > found the alternate gateway. This still leaves me with the mesh node > attached the the ADSL which will not use the other gateway (cos it > thinks it still has one!). > > Is there neat way of forcing a switch? > Has anyone found a neat way of switching internet gateways? > > I have a scenario today whereby an ADSL router has failed. I wanted to > switch to using the other gateway (satellite), which is hung off another > mesh node. > > I tried just rebooting the mesh node with the ADSL router, but all the > other nodes just waited until it came back up again! I tried unplugging > the router, but the mesh node still thought it was a gateway. > > In the end I had to power off the mesh node and then all other nodes > found the alternate gateway. This still leaves me with the mesh node > attached the the ADSL which will not use the other gateway (cos it > thinks it still has one!). > > Is there neat way of forcing a switch? We have a very dense 11 node mesh here along the Columbia River in SW Washington providing free public access wireless to visitors as a city promotional tool. We have been using the Qorvus Qcode mesh management software now for a number of months as I have posted to the list before. Many of the problems discussed on the list associated with mesh management via Wiana and SSH (such as "block Node") have been dealt with in the Qorvus software. I highly recommend anyone working with meshAP deployment take the time to look at their product to see if it will save you the kind of time and frustration it did us. The question really is... what do you want to be doing - dicking around with frustrating and sometimes broken code or expanding and promoting you mesh? BTW - We paid for all our Qorvus product in case anyone thinks I have an ulterior motive for promoting them. > I am seeing the exact same problem with my network. It was not much of > a problem at first but has become a major problem as the network has > grown (26 active nodes now). This would not be so much of a problem if > there was an easier way to block unwanted nodes automatically right now > every time a node reboots I have to manually shell in and block unwanted > routes. My current solution is to put in another gateway on a separate > channel and different ssid and segregating the network to help eliminate > some of the work. I am currently on dev 90 but have had this same > problem on dev 81 85 and 88 I did not have this problem on dev 76 but > the network was much smaller back then too. I may try > reverting everything back to dev 76 before the line for the new > gateway goes in next week. I will keep you posted if I find a GOOD > solution to this problem. > I'm having a performance problem with some of my repeater nodes > after the second hop. I'm using Dev 88. I clear all blocknodes and > let the MeshAP mesh. I then create blocknodes for MeshAP that have > poor quality (both MeshAP gets a blocknode to each other.) This, > believe or not, stabilizes the network so that I can ping and ssh > between each MeshAP. > > I having a problem with some of my repeater getting messages every > *minute*, such as: > > Broadcast Message from root@meshbox <mailto:root@meshbox> > (somewhere) at 19:04 ... > > Attempting to connect gateway 1.x.x.x > > New gateway found at 172.x.x.x > > New gateway found at 172.x.x.x > > Attempting to connect gateway 1.x.x.x > > Attempting to connect gateway 1.x.x.x > > after each attempting to connect message I check reporter and I also > get a new message "*Searching for a gateway to use..."* . > This happens after every attempt to connect to gateway. > > When i try to do a *leechtest* either the performance is extremely > slow or it doesn't work; however, I can ssh into the MeshAP and the > ping time are actually very good. I have extremely strong radio > signals between each MeshAP. > > In Wiana I'm not sure what constitutes a broadcast into noisy or > clean but I have issue a command change the fragmentation "*iwconfig > wlan0 frag 1000*" which is currently not supported by Wiana. > > Wiana has indicated over 30 different AP so I figure it would be > noisy? I'm currently on Channel 1 and all of the different AP are > either on 6 or higher. > > I have remove the preferred gateway setting in Wiana to see if there > is much difference. The answer is it doesn't make that much > difference in my environment. > > When I ssh into distant gateway I also get a hiccup. The screen > freezes momentarily then comes back. This only happens on distant > nodes. > > Unfortunately to get the network stabilized I have to make it > linear. Such that A is the gateway. First my nodes are represented > by A,B,C, and D. There topology is linear. A is the gateway MeshAP > and D the last repeater. > > A Links to B which links to C then to D. > > C and A can see each other but I have stabilized this network by > using a combination of blocknode on A and C ensuring that C goes > through B then to A. > > Also D has blocknode to 2 other MeshAP ensuring that it goes through C. > > Ping time are very good. 10 -14 ms from D to A > > leechtest on A and B is good. > > leechtest on C and D is bad. > > On Each MeshAP: > - Signal strength is excellent. > - Line of site is excellent > - Fresnel Zone is excellent > > A - 200 mW with 8 dbi antenna > B- 100 mW with 8 dbi antenna > C - 200 mW with 8 dbi antenna with 10 degree down tilt built into > the antenna > D - 200 mW with 8 dbi antenna > I have two other nodes E and F that can mesh with B, and D. I > currently using blocknode to prevent them from meshing! > > I have asked a number of people to help me with this problem and it > seems that many of us are having something similar. > > Is this a problem with dev 88? or do other dev have the same problem? > > I guess I finding it frustrating when I deploy a MeshAP that it > getting such problems. I always think it a hardware or placement > problem. > > Is anyone having the same experience? However, I have also checked my XP Pro machine, and this has the same subnet mask of 255.255.255.255 - it does have a default gateway, however. As this is running under DHCP, isn't the mesh router supposed to provide the correct details to clients?! It's an official LW box running dev88. (I can't get dev90 up on it). > eth1 Link encap:Ethernet HWaddr 00:09:5B:91:B1:5A > inet addr:192.168.169.210 Bcast:192.168.169.210 > Mask:255.255.255.255 Your subnet mask is b0rked. It should be 255.255.255.0. # ifconfig eth1 netmask 255.255.255.0 should fix it. Also you'll need a default gateway: # route add default gw 192.168.169.1
}}}