Posted: Fri Sep 02, 2011 1:38 Post subject: Funny thing that WDS
I am having a strange issue
I turn on WDS on three devices
I turn on STP on all three
If I configure for a full mesh (each has an active neighbor with the other two, for failover)
Then wireless DHCP is flaky
sometimes i get an IP, sometimes I don't
Wired ports on the three routers get DHCP from the Main Router every time.
Main Router running the only DHCP Server
others set for DHCP-Server and Disabled
However, I get the same effect if i set them for DHCP-Forwarder with the IP address of the Main Router
I can disable one of the WDS neighbors in the config.
I watch the br0 topology change and congeal again.
(without the possibility of a loop)
and Wireless DHCP is working fine again.
WDS configured all with the same channel, 40Mhz, same SSID, same WPA2-AES-PSK, and as I said, she seems to be perfect until I add the other neighbor and let spanning tree get involved.
I have seen the notes about STP on with WDS and WPA2
and STP off with WDS and WEP or no security.
I assume the presence or absence of STP is just for loopback prevention in this instance?
Main Router
Asus RT-N16
running
dd-wrt.v24-14929_NEWD-2_K2.6_big
(which calls itself mega when on the router)
The other two are
Linksys WRT-310Nv1's
running
dd-wrt.v24-14929_NEWD-2_K2.6_std_nokaid_small
dont use a mesh topology with WDS, it's not supposed to worked that way. if you want to use a mesh, use OLSR or BATMAN _________________ Router: WNDR3300 (wl0: n-Only 5Ghz, WPA2-AES, wl1: g-Only, WPA-Mixed-Mixed)
WDS Node 1: WNDR3300 (wl0: n-Only 5Ghz, WPA2-AES, WDS-connected Router, wl1: g-Only WPA-Mixed-Mixed)
WDS Node 2: WRT54GL (g-Only, WPA-Mixed-Mixed WDS-connected to Router)
Modem: Cisco EPC3202
clients: Notebook 1, D-Link 323, PS3 Slim, Kathrein UFC960 connected to WDS Node 1 via Gigabit Switch. Notebook 2, Deskjet 6980 connected to WDS Node 2
I appreciate that,
But Spanning Tree -is- supposed to act a particular way. And you are right, it isn't working well on WDS.
That is the salient bit of the post.
Why does Spanning Tree -not- work on WDS.
EDIT
ssh to the MainRouter
and the following commands are helpful in inspecting the bridge and spanning tree instance.
brctl show
brctl showstp br0
I moved my root bridge from one of the 310N's to my ASUS
MainRouter# brctl setbridgeprio br0 99
looked up some notes on brctl on the internet
found only one default ddwrt value that deviates from the recommended
forward delay
is set to 1 second by default,
internet indicates it should be set to 4 seconds
to allow for time to coordinate with other devices prior to joining a STP domain
left it alone this time.
I went and changed my hello from 2 seconds(default) to 1 second to get STP hellos more often
brctl sethello br0 1
I went and changed my max message age from 20 seconds (default) to 4 seconds again to get STP to be a bit more proactive on faults.
brctl setmaxage br0 4
wds interfaces appear to the bridge as ethernet and all ethernet get a default 100 cost. no reliability or speed calculations in STP. So I have to find my WDS interface I want to use as a backup
nvram show | grep wds[123]_ | sort
and you get descriptions and MAC addresses which will let you distinguish one WDS router from the other
lines we are looking for are
wl0_wds2_if=wds0.1
wl0_wds3_if=wds0.2
my wds0.2 is the backup link (it has a smaller SNR and lower connection speed per my eyeball)
so i give it a 300 cost on both the Main and Media Routers
Yes actually, I will be happy to add this info. I use WDS and STP and your findings will be helpful. If you could post the exact startup script then I can copy / paste it in. ( I'm a copy / paste level Linix Guru ). If you want to post a complete transcript to add that would earn you several extra beers. _________________ http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Posted: Sat Sep 03, 2011 23:32 Post subject: Write up with some references
First let me apologize for the lack of actual art
I am an ASCII art type guy. Feel free to replace as you see fit.
Now the initial state:
Three routers
Main MAC:xxxxxxxxxxxx Asus RT-N16 running dd-wrt.v24-14929_NEWD-2_K2.6_big
Media MAC:xxxxxxxxxxx1 WRT-310Nv1 running dd-wrt.v24-14929_NEWD-2_K2.6_std_nokaid_small
Mid MAC:xxxxxxxxxxx2 WRT-310Nv1 running dd-wrt.v24-14929_NEWD-2_K2.6_std_nokaid_small
Main Has the WAN link to a PPPoE DSL
Link1 = WDS neighbors Main-Mid
Link2 = WDS neighbors Mid-Media
Link3 = WDS neighbors Media-Main
By default DD-WRT exposes a single button on the webpage to toggle Spanning Tree On/Off
I was having a problem with DHCP and topology changes on Spanning Tree and suspected that STP was not working properly.
Specifically, my client would associate with the AP(dd-wrt router) but not get a DHCP assigned address.
Curiously the wired ports on the back seemed to always work just fine.
The issue was intermittant and bugged me greatly, in that I could simply turn off Link3 and have no issues at all for months at a time.
Since it was based on Linux
I googled up spanning tree and ubuntu and linux and various combinations and found that Linux has been using brctl since kernel 2.2.
I found that none of the controls were exposed in the DD-WRT front end, so I ssh'd to the devices.
(feel free to telnet or use the run command on the user interface for what follows)
first I checked to see if brctl was available with this command
Code:
root@Main:~#which brctl
/usr/sbin/brctl
I got it!!
now to business
brctl comes with the following options I will explain in a moment
Code:
Usage: brctl [commands]
commands:
addbr <bridge> add bridge
delbr <bridge> delete bridge
addif <bridge> <device> add interface to bridge
delif <bridge> <device> delete interface from bridge
setageing <bridge> <time> set ageing time
setbridgeprio <bridge> <prio> set bridge priority
setfd <bridge> <time> set bridge forward delay
sethello <bridge> <time> set hello time
setmaxage <bridge> <time> set max message age
setpathcost <bridge> <port> <cost> set path cost
setportprio <bridge> <port> <prio> set port priority
setportsnooping <bridge> <port> <addr> set port snooping
clearportsnooping <bridge> <port> <addr> clear port snooping
showportsnooping <bridge> show port snooping
enableportsnooping <enable> enable port snooping
show show a list of bridges
showmacs <bridge> show a list of mac addrs
showstp <bridge> show bridge stp info
stp <bridge> {on|off} turn stp on/off
I peeked around to see what bridges were available
Code:
root@Main:~# brctl show
bridge name bridge id STP enabled interfaces
br0 8000.xxxxxxxxxxxx yes vlan1
eth1
wds0.1
wds0.2
Which tells me I have one bridge named br0
Code:
root@Main:~# brctl showstp br0
br0
bridge id 8000.xxxxxxxxxxxx
designated root 8000.xxxxxxxxxxx1
root port 3 path cost 100
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 0.00 tcn timer 1.79
topology change timer 0.00 gc timer 0.83
flags TOPOLOGY_CHANGE_DETECTED
vlan1 (1)
port id 8001 state forwarding
designated root 8000.xxxxxxxxxxx1 path cost 100
designated bridge 8000.xxxxxxxxxxxx message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 100 hold timer 0.83
flags
eth1 (2)
port id 8002 state forwarding
designated root 8000.xxxxxxxxxxx1 path cost 100
designated bridge 8000.xxxxxxxxxxxx message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 100 hold timer 0.83
flags
wds0.1 (3)
port id 8003 state forwarding
designated root 8000.xxxxxxxxxxx1 path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 19.85
designated port 8004 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
wds0.2 (4)
port id 8004 state forwarding
designated root 8000.xxxxxxxxxxx1 path cost 100
designated bridge 8000.xxxxxxxxxxxx message age timer 0.00
designated port 8004 forward delay timer 0.00
designated cost 100 hold timer 0.83
flags
root@Main:~#
Let's take a moment to interpret what we are seeing
bridge id is the MAC of your local bridge interface with "8000." out in front (8000.xxxxxxxxxxxx)
designated root is the MAC of the root bridge (spanning tree boss) with "8000." out in front (8000.xxxxxxxxxxx1)
root port is the port used to communicate with the root bridge (this case 3 which is wds0.1)
path cost is the culmulative cost of the path to root (how far away you are from the root bridge)
max age and bridge max age are the maximum age of the STP topology which is 20 seconds by default
hello time and bridge hello time is the number of seconds between hello messages on the bridge
forward delay and bridge forward delay are the time STP will wait before forwarding packets onto an existing bridge
which gives the new device time to find out what is going on before she starts talking into a network
ageing time is the length of time a MAC address is held in the forwarding table
the forwarding table remembers the source MAC when a packet crosses,
then it will store that MAC/port relation so when an packet comes in destined for the MAC
the switch will then forward the packet back down the port where it is last known to be
hello timer, tcn timer, topology change timer, and gc timer are all the current values in the live timers
flags is again a live value containing the last active flag value (my example indicates a topology change has been detected)
Looking at the remaining info are the port names(vlan1, eth1, wds0.1, wds0.2), physical port (number following the port names "1, 2, 3, 4"), port ids (8001-8004),
state (forwarding), path cost (all ethernet ports are default 100, we will revisit this in a moment) , and a bunch of timers and stuff.
The root bridge appears to be my Mid Router
and the path costs are default all the way around.
We would repeat the above procedure on the other two devices to ensure they all believed the root was the Mid Router and that all the path costs were the same.
in my case they were. So STP should have calculated a distance to the root bridge via link3 to be 200 (path cost of link3 added to the path cost of link 2).
Which it did not do in this case, because I initially had link3 disabled on the Media router. (so it would work)
first thing I wanted to do and I suggest for all involved.
Bring the root bridge to the WAN router, Main Router, in this case the newest and most powerful device ASUS RT-N16
By default, DDWRT assigns a root bridge priority of 100, I want my Main Router to be root so I influence this decision by lowering the Main Routers bridge priority to 99.
NOTE; real switch vendors set their default bridge priority much higher than this, cisco and brocade uses 32768 out of the box. There are some other out of the box default settings that DDWRT may want to review for compatibility with industry standard commercial hardware. Priority is one of the "needs to play nice" ones where the default 100 will make the DDWRT device assume root bridge from an existing STP domain.
http://www.brocade.com/support/Product_Manuals/ServerIron_SwitchRouterGuide/STP.pdf http://www.cisco.com/en/US/docs/switches/lan/catalyst4500/12.1/8aew/configuration/guide/spantree.html#wp1020334
you cant wrap code in color, first actual change is to move the root where I want it
Code:
root@Main:~# brctl setbridgeprio br0 99
Code:
root@Main:~# brctl showstp br0
br0
bridge id 0063.xxxxxxxxxxxx
designated root 0063.xxxxxxxxxxxx
root port 0 path cost 0
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 1.81 tcn timer 0.00
topology change timer 0.00 gc timer 0.81
flags
vlan1 (1)
port id 8001 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 0 hold timer 0.81
flags
eth1 (2)
port id 8002 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 0 hold timer 0.81
flags
wds0.1 (3)
port id 8003 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8003 forward delay timer 0.00
designated cost 0 hold timer 0.81
flags
wds0.2 (4)
port id 8004 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8004 forward delay timer 0.00
designated cost 0 hold timer 0.81
flags
root@Main:~#
root@Mid:~# brctl showstp br0
br0
bridge id 8000.xxxxxxxxxxx1
designated root 0063.xxxxxxxxxxxx
root port 4 path cost 100
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 0.00 tcn timer 0.00
topology change timer 0.00 gc timer 1.24
flags
eth0 (1)
port id 8001 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 100 hold timer 0.23
flags
vlan1 (2)
port id 8002 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 100 hold timer 0.23
flags
wds0.1 (3)
port id 8003 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 0.00
designated port 8003 forward delay timer 0.00
designated cost 100 hold timer 0.23
flags
wds0.2 (4)
port id 8004 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 19.20
designated port 8003 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
root@Mid:~#
root@Media:~# brctl showstp br0
br0
bridge id 8000.xxxxxxxxxxx2
designated root 0063.xxxxxxxxxxxx
root port 3 path cost 200
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 0.00 tcn timer 0.12
topology change timer 0.00 gc timer 1.08
flags TOPOLOGY_CHANGE_DETECTED
eth0 (1)
port id 8001 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx2 message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 200 hold timer 1.08
flags
vlan1 (2)
port id 8002 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx2 message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 200 hold timer 1.08
flags
wds0.1 (3)
port id 8003 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 19.36
designated port 8003 forward delay timer 0.00
designated cost 100 hold timer 0.00
flags
root@Media:~#
And we can now see that the ASUS xxxxxxxxxxxx is the designated root across all three devices
and the path cost to get to the root from the Media router is 200 (link1 plus link2)
now I want link3 to be less preferred so I want to make it's cost higher than the natural 100 (cost of link1 directly back to the root)
but first i have to figure out which link is which.. hmm. that is stored in the nvram variables
I see that the one I am interested in is wds0.2
so my command is
My second change was to make the cost of link three higher than the cost of link1 plus link2 (201) would have worked to but I am big on round numbers
This is the first of two, I adjust the cost it on both sides of the connection,because I am a symmetric type of guy
Code:
root@Main:~#brctl setpathcost br0 wds0.2 300
Code:
root@Main:~# brctl showstp br0
br0
bridge id 0063.xxxxxxxxxxxx
designated root 0063.xxxxxxxxxxxx
root port 0 path cost 0
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 0.69 tcn timer 0.00
topology change timer 0.00 gc timer 0.69
flags
vlan1 (1)
port id 8001 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
eth1 (2)
port id 8002 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
wds0.1 (3)
port id 8003 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8003 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
wds0.2 (4)
port id 8004 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 300
designated bridge 0063.xxxxxxxxxxxx message age timer 0.00
designated port 8004 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
root@Main:~#
now I have to enable the WDS port in the Media router so I can see it's port name
so we do that in the GUI
repeat the
and we see it is wds0.1 now.
order is important this value will change based on where on the list it lies
be careful when adding and disabling values on the webpage, these will move about and may cause unintended side effects.
Here we cost the Media side of the Media-Main Link3
Code:
root@Media:~#brctl setpathcost br0 wds0.1 300
Code:
root@Media:~# brctl showstp br0
br0
bridge id 8000.xxxxxxxxxxx2
designated root 0063.xxxxxxxxxxxx
root port 4 path cost 200
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 1.00 bridge forward delay 1.00
ageing time 300.00
hello timer 0.00 tcn timer 0.04
topology change timer 0.00 gc timer 105.30
flags
eth0 (1)
port id 8001 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx2 message age timer 0.00
designated port 8001 forward delay timer 0.00
designated cost 200 hold timer 0.66
flags
vlan1 (2)
port id 8002 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx2 message age timer 0.00
designated port 8002 forward delay timer 0.00
designated cost 200 hold timer 0.66
flags
wds0.1 (3)
port id 8003 state blocking
designated root 0063.xxxxxxxxxxxx path cost 300
designated bridge 0063.xxxxxxxxxxxx message age timer 3.41
designated port 8004 forward delay timer 0.00
designated cost 0 hold timer 0.00
flags
wds0.2 (4)
port id 8004 state forwarding
designated root 0063.xxxxxxxxxxxx path cost 100
designated bridge 8000.xxxxxxxxxxx1 message age timer 2.40
designated port 8003 forward delay timer 0.00
designated cost 100 hold timer 0.00
flags
setageing <bridge> <time> set ageing time (Default 300 seconds)
Sets the aging time. The aging time is the number of seconds a MAC-address will be kept in the forwarding database after having received a packet from this MAC address. The entries in the forwarding database are periodically timed out to ensure they won't stay around forever. Normally there should be no need to modify this parameter.
sets the ethernet (MAC) address ageing time, in seconds. After <time> seconds of not having seen a frame coming from a certain address, the bridge will time out (delete) that address from the Forwarding DataBase (fdb).
setbridgeprio <bridge> <prio> set bridge priority (default 100, lower is more preferred, industry standard default is 32768)
Sets the bridge's relative priority. The bridge with the lowest priority will be elected as the root bridge. The root bridge is the "central" bridge in the spanning tree.
sets the bridge's priority to <priority>. The priority value is an unsigned 16-bit quantity (a number between 0 and 65535), and has no dimension. Lower priority values are 'better'. The bridge with the lowest priority will be elected 'root bridge'.
setfd <bridge> <time> set bridge forward delay (default is 1, should be 4)
Sets the forwarding delay time. The forwarding delay time is the time spent in each of the Listening and Learning states before the Forwarding state is entered.
sets the bridge's 'bridge forward delay' to <time> seconds
sethello <bridge> <time> set hello time (default is 2, can be 1)
Sets the hello time. Every (this number) seconds, a hello packet is sent out by the Root Bridge and the Designated Bridges. Hello packets are used to communicate information about the topology throughout the entire Bridged Local Area Network.
sets the bridge's 'bridge hello time' to <time> seconds
setmaxage <bridge> <time> set max message age (default is 20, w/hello at 1, can be 4)
Sets the maximum message age. If the last seen (received) hello packet is more than this number of seconds old, the bridge in question will start the takeover procedure in attempt to become the Root Bridge itself.
sets the bridge's 'maximum message age' to <time> seconds
setpathcost <bridge> <port> <cost> set path cost (default is 100, higher is less preferred)
Sets the cost of receiving (or sending, I'm not sure) a packet on this interface. Faster interfaces should have lower path costs. These values are used in the computation of the minimal spanning tree. Paths with lower costs are likelier to be used in the spanning tree than high-cost paths (As an example, think of a gigabit line with a 100Mbit or 10Mbit line as a backup line. You don't want the 10/100Mbit line to become the primary line there.)
The Linux implementation currently sets the path cost of all eth* interfaces to 100, the nominal cost for a 10Mbit connection. There is unfortunately no easy way to discern 10Mbit from 100Mbit from 1Gbit Ethernet cards, so the bridge cannot use the real interface speed.
sets the port cost of the port <port> to <cost>. This is a dimensionless metric.
setportprio <bridge> <port> <prio> set port priority
sets the port <port>'s priority to <priority>. The priority value is an unsigned 8-bit quantity (a number between 0 and 255), and has no dimension. This metric is used in the designated port and root port selection algorithms. You use this one to distinguish between two equal cost ports.
To get mine to be stable and responsive to topology changes
I tweaked it as follows
I set the Hello to every one second vs two seconds, to make it detect a topology change faster
I changed the maximum age from twenty seconds to four seconds to make it refresh the topology more often.
I set the Forwarding delay to four seconds vs one second, to make it wait longer before participating in a STP domain
in the startup commands I put the following
Saved each,
then rebooted Main, Mid, and Media in sequence to make sure all were working from the same instructions
hopped back in on the command line
and looked at the spanning tree to ensure I was seeing what I expected.
Now she is up and stable
to test
i took a wireless client from on the Media router
started a streaming ping to google
removed power from the Mid router
she took about 30 seconds to switch paths
restored power to the Mid router
long pause while she booted while the ping was successful
then ping stopped for about 2 mins
then restored via the Mid router
removed power again from Mid router
ping stopped for about 2 mins
then restored over link3
restored power again to the Mid router
long pause while she booted while the ping was successful
then ping stopped for about 30 seconds
then restored via the Mid router
not exactly a scientific test but she sufficed for my purposes.
I can fail to an alternate path in less than 3 minutes if one of my routers go down.
Which is really good for consumer applications.
Last edited by nariub on Sun Sep 04, 2011 15:09; edited 1 time in total
Joined: 06 Jun 2006 Posts: 3763 Location: I'm the one on the plate.
Posted: Sun Sep 04, 2011 1:21 Post subject:
Do you think this should go in the STP Wiki, or maybe a new entry under WDS ? "WDS Mesh Script" perhaps ? _________________ http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Spanning Tree is Spanning Tree.
It's where I looked for information on how to query STP on the device to find out what was going on.
you could put a note on the WDS pages regarding MESH-> see STP.. because MESH WDS is really gonna need an understanding of STP.
but this note illustrates an investigation and configuration of STP via tools not available in the UserInterface.
if i had my druthers,
i would have liked root bridge priorty default set to industry standards and exposed on the routing/bridge tab. forward delay and others would be nice too.
I would also like the path cost exposed on the WDS page so the user interface would work out which is wds0.1 and which is wds0.2, etc because my way seems a little horky if i have several in there and choose to disable one high on the list. I now gotta go in and bump all the interface costs down one.
but the picture is fine. large but fine.
Again it is an exercise in command line STP, Applicable to other stuff than just WDS.
Joined: 06 Jun 2006 Posts: 3763 Location: I'm the one on the plate.
Posted: Sun Sep 04, 2011 5:01 Post subject:
OK, it is posted. I couldn't get my artwork to take, but your stick diagram looks better anyway. Let me know of any changes it might need. _________________ http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Just checked, couldnt find it under "WDS Mesh". I should probably wait till tomorrow to see it posted.
Will try it out and report on findings in a e3000 and WRT610n hybird environment. _________________ ===================================
1 * DIR-866L - 29193 Mega (Main Gateway)
1 * EA4200 - 29193 Mega (Main Gateway)
1 * EA6500 - 29193 Mega (Repeater Bridge)
1 * EA6500v2 - 29193 Mega (Repeater Bridge)
1 * WRT610N - 29193 Mega (Repeater Bridge)
===================================
Joined: 06 Jun 2006 Posts: 3763 Location: I'm the one on the plate.
Posted: Sun Sep 04, 2011 5:17 Post subject:
Do the commands ( not the scripts ) need an NVram commit to make them survive a reboot ?
ndewan wrote:
Just checked, couldnt find it under "WDS Mesh".
I can't seem to edit any menus without screwing something else up. The STP entry should suffice.
Edit -
Also, could we label each of the actual changes with a step number ? For the mentally challenged such as myself it may not be clear which data is effecting system changes versus what is informational. If you want, just edit your post to make the commands that change something, and the start up scripts, to be red text and then I'll add step numbers to the WiKi.
I'll also fix the BOLD text tomorrow. The HTML did not survive the copy / paste. _________________ http://69.175.13.131:8015 Streaming Week-End Disco. Station Ripper V 1.1 will do.
Generalized
if you have less than three devices,
the default options should suffice
if you have three or more
query the native state of the bridge on each device, this will tell you what STP wants to do if left to its own devices.
I then suggest
Move the root bridge to the WAN facing router
(and add the command to the startup to make it permanent)
query the bridge on each device, to ensure they all still agree on who the root bridge is.
now it gets progressively more complicated
remember three devices equals three links
four devices will be six links
five devices will be ten links
Make a diagram of the devices and links
stare at it and try to find a loop free topology on your diagram
Failure analysis
now imagine if any one of your devices fail, what is the impact on your topology, what alternative path do you want it to use.
depending on how retentive you are,
repeat for each single device failure
repeat for any two devices, any three, etc until you get below two devices. at two devices we revert to it doesnt really matter that much any
since none of these are physically wired links, you dont have to repeat the above procedure for each defined link, or 2links or 3links, etc
based on the failure analysis
create a path cost plan that will support your topology and failover plan. (anyone see the advantage of small spanning tree domains yet?)
Remember you can you port cost or port priority
i am symmetrical about port cost, adding the same cost to both side of the link
port priority is used when links are the same cost it prefers one over the other. default is zero
my example above,
i could have simply changed port priority of link3 on the Media router if that was the only criteria. but the cost to the root via link3 was 100, the cost to the root via link2 was 200. In order to get priority to work, i would have had to recost link3 to 200 -and- set the priority to 1. I am a minimalist in this regard, why make two changes when one will suffice.
technically these are all locally significant, meaning i did not have to set the port cost on the root bridge (main).. i did because i have an addiction to symmetry. it is a personal issue i have battled all my life.
Enable all the WDS links on one device at a time, (this will allow you to use the nvram commands to see which wds0.X each of these are.
Write this down
Disable all the WDS links before leaving the device
define your plan,
put your commands for link costing in the startup config (admin > commands > startup) save config
(if you want to change hello/fwd delay/mess age you need to do this now too)
reboot all your devices
Enable all the WDS links on each device, and reboot
as they come up from the second reboot
they will come up with the WDS costing and hopefully the topology you desire
Query the devices to see if your settings took
Perform testing
based on your Failure analysis
fail each device, 2 devices, 3 devices etc
to see if she recovers on the backup path(s)
This also serves to tell you what your baseline recovery times are.
save your drawings/notes, device configs, log files etc. So you can rebuild it when sunspots destroy all your household electronics that survived the zombie attacks
