WRT54G (v4) Client mode

I asked a question along these lines a few months back and
was instructed how to (with DD-WRT firmware installed)
do a wireless survey and then "join" another WAP's network.

I was also warned (by whom I forgot) that doing this would
"screw up" my WRT54G so I never did it.

What I want to do is mount my WRT54G on a tall tower in
a weatherproof enclosure. I would then (preferably wirelessly)
connect with the router and use it as a client mode repeater
to connect with a nearby free AP.

Can this be done?
--
73's de Ken KG0WX - Kadiddlehopper #11808,
Flying Pigs #-1055, Grid EM17io, FT-857D,
Elecraft XG2, 4SQRP Tenna Dipper, Heath GD-1B

"Ken Bessler" <(E-Mail Removed)> hath wroth:

>I asked a question along these lines a few months back and
>was instructed how to (with DD-WRT firmware installed)
>do a wireless survey and then "join" another WAP's network.
>
>I was also warned (by whom I forgot) that doing this would
>"screw up" my WRT54G so I never did it.

Probably me. We're both apparently too lazy to Google for my original
posting so I'll explain it again.

If your WRT54G running DD-WRT is set to the access point mode, then
hitting "join" in the site survey page will cause it to:
1. Switch to the client mode.
2. Change a few settings for no obvious reason. I forgot exactly
which ones but I think one was to disable the DHCP server.
3. Automagically save the settings. (this may have changed in later
releases of the firmware).

If you want to go back to access point mode after doing this, you're
stuck with finding all the things that changed, putting eveything
back, and saving the results.

The way around this is to first save your current working
configuration to a file. Do the join thing or whatever. When you're
done, restore the settings from the file. In other words, make a
backup first.

>What I want to do is mount my WRT54G on a tall tower in
>a weatherproof enclosure. I would then (preferably wirelessly)
>connect with the router and use it as a client mode repeater
>to connect with a nearby free AP.

There's no such thing as a "client mode repeater". It's either:
1. access point
2. client mode
3. repeater

What you're describing is a repeater. It should work as described
including hitting the join button. As long as you can get to the IP
address of the repeater for configuration, you should be fine.

>Can this be done?

Yes. Make sure you have the final version of V23 SP1 installed.
Too many bugs in earlier versions.

Incidentally, if you can run power up your tall tower, then you can
run CAT5 cable for PoE (power over ethernet) to the WRT54G. The
WRT54G has a wide range switching regulator that will run on anything
from about 4VDC to perhaps 18VDC. Just take the spare 4 wires in the
CAT5 cable, connect one end to the power plug on the WRT54G, and the
other end to the 12v 1A wall wart. There will be some cable
resistance losses, but it will still work just fine. Then, you can
forget about using it as a repeater (with all it's complications) and
just plug your computer into the CAT5 connections. The WRT54G client
mode can handle multiple computers, so you don't need a router.


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Great info here.

- So when in Client mode, DHCP is disabled.

- If another WRT54G is placed right behind the client and wired in, can the
second WRT54G then act as the Wireless AP / DHCP server? I'm assuming yes,
but please confirm.



"Jeff Liebermann" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> "Ken Bessler" <(E-Mail Removed)> hath wroth:
>
>>I asked a question along these lines a few months back and
>>was instructed how to (with DD-WRT firmware installed)
>>do a wireless survey and then "join" another WAP's network.
>>
>>I was also warned (by whom I forgot) that doing this would
>>"screw up" my WRT54G so I never did it.
>
> Probably me. We're both apparently too lazy to Google for my original
> posting so I'll explain it again.
>
> If your WRT54G running DD-WRT is set to the access point mode, then
> hitting "join" in the site survey page will cause it to:
> 1. Switch to the client mode.
> 2. Change a few settings for no obvious reason. I forgot exactly
> which ones but I think one was to disable the DHCP server.
> 3. Automagically save the settings. (this may have changed in later
> releases of the firmware).
>
> If you want to go back to access point mode after doing this, you're
> stuck with finding all the things that changed, putting eveything
> back, and saving the results.
>
> The way around this is to first save your current working
> configuration to a file. Do the join thing or whatever. When you're
> done, restore the settings from the file. In other words, make a
> backup first.
>
>>What I want to do is mount my WRT54G on a tall tower in
>>a weatherproof enclosure. I would then (preferably wirelessly)
>>connect with the router and use it as a client mode repeater
>>to connect with a nearby free AP.
>
> There's no such thing as a "client mode repeater". It's either:
> 1. access point
> 2. client mode
> 3. repeater
>
> What you're describing is a repeater. It should work as described
> including hitting the join button. As long as you can get to the IP
> address of the repeater for configuration, you should be fine.
>
>>Can this be done?
>
> Yes. Make sure you have the final version of V23 SP1 installed.
> Too many bugs in earlier versions.
>
> Incidentally, if you can run power up your tall tower, then you can
> run CAT5 cable for PoE (power over ethernet) to the WRT54G. The
> WRT54G has a wide range switching regulator that will run on anything
> from about 4VDC to perhaps 18VDC. Just take the spare 4 wires in the
> CAT5 cable, connect one end to the power plug on the WRT54G, and the
> other end to the 12v 1A wall wart. There will be some cable
> resistance losses, but it will still work just fine. Then, you can
> forget about using it as a repeater (with all it's complications) and
> just plug your computer into the CAT5 connections. The WRT54G client
> mode can handle multiple computers, so you don't need a router.
>
>
> --
> Jeff Liebermann (E-Mail Removed)
> 150 Felker St #D http://www.LearnByDestroying.com
> Santa Cruz CA 95060 http://802.11junk.com
> Skype: JeffLiebermann AE6KS 831-336-2558

"TH" <(E-Mail Removed)> hath wroth:

>- So when in Client mode, DHCP is disabled.

I'm not sure, but that's what I recall. I don't want to mess with my
WRT54G at this time to verify it. However, if it doesn't disable the
DHCP server, you should do it yourself. It's a bad idea to have two
DHCP servers running. (One at the ISP and one in your client radio or
repeater).

>- If another WRT54G is placed right behind the client and wired in, can the
>second WRT54G then act as the Wireless AP / DHCP server? I'm assuming yes,
>but please confirm.

Yep. However, watch the terminology please. A wireless access point
does not have a router to do NAT. Therefore, that will just act as a
repeater for the wireless client. Access points normally don't have
DHCP servers.

What I think you meant to ask is whether a 2nd WRT54G can act as a
wireless router with DHCP server (and NAT). Yes, that will also work.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Hi, y'all,

Jeff Liebermann wrote:
> "Ken Bessler" <(E-Mail Removed)> hath wroth:
>
> >What I want to do is mount my WRT54G on a tall tower in
> >a weatherproof enclosure. I would then (preferably wirelessly)
> >connect with the router and use it as a client mode repeater
> >to connect with a nearby free AP.
>
> There's no such thing as a "client mode repeater". It's either:
> 1. access point
> 2. client mode
> 3. repeater
>
> What you're describing is a repeater. It should work as described
> including hitting the join button. As long as you can get to the IP
> address of the repeater for configuration, you should be fine.
>
> >Can this be done?
>
> Yes. Make sure you have the final version of V23 SP1 installed.
> Too many bugs in earlier versions.

DAYUM! Isn't this what I've been pulling my hair out over???

I know I'm clueless - it's been proven many times in many venues.

But I thought I heard him say he wanted to put something on a pole,
power it, see it with wifi, and use it to get to any AP he could see -
or in this case, a free AP, but I presume that any free one would do.

That's what I've burned an awful lot of electronic trees printing into
hyperspace trying to succeed in doing, all to no avail (at least yet).
That I could also provide a hard wire feed, if I needed it, to
something else (see below) would be a bonus, of course.

>
> Incidentally, if you can run power up your tall tower, then you can
> run CAT5 cable for PoE (power over ethernet) to the WRT54G. The
> WRT54G has a wide range switching regulator that will run on anything
> from about 4VDC to perhaps 18VDC. Just take the spare 4 wires in the
> CAT5 cable, connect one end to the power plug on the WRT54G, and the
> other end to the 12v 1A wall wart. There will be some cable
> resistance losses, but it will still work just fine. Then, you can
> forget about using it as a repeater (with all it's complications) and
> just plug your computer into the CAT5 connections. The WRT54G client
> mode can handle multiple computers, so you don't need a router.

However, I'm not sure what you're saying. With the immediately above,
can I still do this wirelessly, and have an ethernet pigtail for other
devices (a router, a VoIP box, e.g.), or are you saying I have to
connect to it via ethernet in order to make it work (at al)?

I've been confused for a long time, but now I'm more confused...

Hayulp!

L8R

Skip

Morgan 461 #2
SV Flying Pig KI4MPC
http://tinyurl.com/p7rb4 - NOTE:new URL! The vessel as Tehamana, as we
bought her

"Believe me, my young friend, there is *nothing*-absolutely
nothing-half so much worth doing as simply messing,
messing-about-in-boats; messing about in boats-or *with* boats.
In or out of 'em, it doesn't matter. Nothing seems really to matter,
that's the charm of it.
Whether you get away, or whether you don't; whether you arrive at your
destination or whether you reach somewhere else, or whether you never
get anywhere at all, you're always busy, and you never do anything in
particular; and when you've done it there's always something else to
do, and you can do it if you like, but you'd much better not."

I think the point is if you want to have more than one computer *on the
boat* using the wireless link you'd do well to consider a 'dual device'
setup.

One with antennae aimed toward picking up the signal from shore. To this
you'd connect *another* device with wired ethernet. It's this second device
that would provide the wireless signals to the other devices on the the
boat. This is what I'm in the process of doing for our boat.

I'm hoping to let my wife fire up the laptop and use the boat's SSID, not
have to hunt up one from shore. Realizing, of course, that a visit to the
web configuration page for the shore-connecting device might be required to
join up with the shore SSID. But the laptops on the boat won't have to have
any part of their configuration changed. I fully expect, at some point, to
have an on-boat computer doing the babysitting of the shore link. Since the
boat has GPS it should be within the realm of possibilities to have it
automagically tracking what SSIDs are 'appropriate' for use based on
physical locations. As in, remember what access points I've told it to use,
at which locations, and attempt to reconnect without intervention.

What I've done, thus far, is use a pair of WRT54G (both version 2 devices)
and loaded the latest DD-WRT (2.3sp1?) onto them. The one making the shore
connection is running in Client mode. The one providing Wifi to the boat
devices is running as an access point. The on-boat device is connected one
of the four switch ports on the shore-link device. There's nothing
connected into either device's WAN port.

So far it seems to be working, at least within the limits of getting the
shore-link established trying various different antennae.

Note the careful avoidance of terms like router, gateway and the like.
These can be somewhat loaded terms when trying to figure out this sort of
stuff. While the various terms DO end up being appropriate for the tasks
being performed, it can be confusing if they're used in the wrong contexts.
As in, yes, the on-boat device is a "router", in the strictest hardware
sense as it's default linksys firmware provides. The default factory mode
is as a "gateway", but that term also applies at the PC-level as past of the
IP configuration. But I'm using it for neither of those purposes, instead
as a Access Point. Same deal goes for the shore-link device being used as a
'Client'. Thus it can get confusing for someone new to the stuff to get
their head around it all.

If you try to do it with just a single device, with a single radio (two
antenna don't mean two radios) then you're really never going to see
effective performance. That sort of 'repeater' mode basically wastes half
the bandwidth. That and trying to get WDS working is a pain in the ass.
Yes, it "ought" to be able to work, but in practice it just doesn't (yet?).

Now, combine all the above with trying to use other devices besides the
WRT54g and, well, yeah it's going to be a challenge. It's all getting 'less
worse' but it's still a far cry from being easy.

-Bill Kearney

Bill sent me this directly; I hadn't realized he'd sent it here as
well....

Hi, Bill, and thanks for the note.

Comments/followup inline:

>I think the point is if you want to have more than one computer *on the
> boat* using the wireless link you'd do well to consider a 'dual device'
> setup.

As you probably have seen from John Navas' and my conversations, that's
what
I started out with. The two devices were the highly touted (note that I

never have seen a user report - just touts) Senao 2611CB3 Deluxe units.
The
ones I was sent were the ones with no housings - shown as "OD", mostly,

apparently, for use in a NEMA enclosure.

However, I can't get either of the units to work as a wired bridge, and
- at
least based on last night's efforts - they also don't want to do much
else,
either. The one which finally saw all the available clients eventually

stopped responding to my attempts to interrogate it. While I didn't do
a
power down/restore (which I could do, and in fact, since I'm a sailboat
and
won't have them up all the time due to load issues [though in a windy
and
sunny location, I could leave them up during the day, as I have ample
solar
and wind generation]), I can say for sure that until I have something
successfully breadboarded a considerable amount of time, I'm not going
to
put it on top of the mast, where access will be only catastrophic, not
a
"push the button to reboot"!

Then, to add insult to injury, when I try to connect the two together,
IP
conflicts abound. It doesn't matter if they're 6' apart (the supplied
cable) or 5" (my made-up crossover pigtail), it does the same thing.
It
also doesn't matter if there's a router in between, as one
correspondent
went to some considerable detail to suggest would cure the problem.
That's
so whether I use a really offbeat IP and subnet set (such as 15.15.15.1
and
255.0.0.0) which can tell my NIC to use and interrogate and set them up
in
whatever fashion I want. There's no way all the crazy IP sets I used
can
all be causing conflicts with my laptop(s - regardless of which I try
it
in). Yet the vendor is trying to tell me - and Discovercard! - I have
to
engage a MSCertified NAnalyst to resolve the IP conflicts in my
machine...

>
> One with antennae aimed toward picking up the signal from shore. To this
> you'd connect *another* device with wired ethernet. It's this second
> device
> that would provide the wireless signals to the other devices on the the
> boat. This is what I'm in the process of doing for our boat.

Yah, I'd hoped to provide all the cruisers in the harbor with another
AP in
effect, but this time local to the top of my mast. I was going to have
the
stick up top, and the duck below, the NEMA. Vendor assured me that
would
work just fine. Of course, I have since come to regard anything from
him as
both uninformed and dangerous. Many hours on the phone (in the first
few
weeks - he's since refused to talk to me) doing exactly what he said to
do
would not cure the problem(s) - but he refuses to take it back.

>
> I'm hoping to let my wife fire up the laptop and use the boat's SSID, not
> have to hunt up one from shore. Realizing, of course, that a visit to the
> web configuration page for the shore-connecting device might be required
> to
> join up with the shore SSID. But the laptops on the boat won't have to
> have
> any part of their configuration changed. I fully expect, at some point,
> to
> have an on-boat computer doing the babysitting of the shore link. Since
> the
> boat has GPS it should be within the realm of possibilities to have it
> automagically tracking what SSIDs are 'appropriate' for use based on
> physical locations. As in, remember what access points I've told it to
> use,
> at which locations, and attempt to reconnect without intervention.

Cool idea. My cruisng grounds are unlikely to be as repetitive - we'll

either be in one place for a long time, or not back in the short term
(like
perhaps more than a year, in which case there likely will be new ones,
and
better ones, anyway). However, I'm not enthusiastic about causing
another
set of wiring and other challenges as would be required with a
full-time
computer aboard, and as above, so - while I'd far rather use the
equivalent
of the WZC, manually entering the SSID is manageable, so long as I can
get
to the bridge to configure it.

So far, if I have the bridge in dhcp, it's invisible for configuration,
and
nothing will change it other than a hard reboot; power cycling doesn't
change it, we have to go back to factory default to get it out of dhcp.

That clearly won't work up the mast. Not up the mast makes for no
signal if
I were to try to use coax...


>
> What I've done, thus far, is use a pair of WRT54G (both version 2 devices)
> and loaded the latest DD-WRT (2.3sp1?) onto them. The one making the
> shore
> connection is running in Client mode. The one providing Wifi to the boat
> devices is running as an access point. The on-boat device is connected
> one
> of the four switch ports on the shore-link device. There's nothing
> connected into either device's WAN port.
>
> So far it seems to be working, at least within the limits of getting the
> shore-link established trying various different antennae.

Cool. However, finding the early versions, I gather, will be very
challenging, and from what I read, the current versions don't perform
at all
like what we need them to do. I *would* like more power, on the
premise
that reaching the shore would need it (the higher gain antenna
supposedly
takes care of the hearing part).

>
> Note the careful avoidance of terms like router, gateway and the like.
> These can be somewhat loaded terms when trying to figure out this sort of
> stuff. While the various terms DO end up being appropriate for the tasks
> being performed, it can be confusing if they're used in the wrong
> contexts.
> As in, yes, the on-boat device is a "router", in the strictest hardware
> sense as it's default linksys firmware provides. The default factory mode
> is as a "gateway", but that term also applies at the PC-level as past of
> the
> IP configuration. But I'm using it for neither of those purposes, instead
> as a Access Point. Same deal goes for the shore-link device being used as
> a
> 'Client'. Thus it can get confusing for someone new to the stuff to get
> their head around it all.

Heh. I don't know if I'm being confused, or merely stymied by inop
gear...

>
> If you try to do it with just a single device, with a single radio (two
> antenna don't mean two radios) then you're really never going to see
> effective performance. That sort of 'repeater' mode basically wastes
> half
> the bandwidth. That and trying to get WDS working is a pain in the ass.
> Yes, it "ought" to be able to work, but in practice it just doesn't
> (yet?).

I don't mind two pieces of gear - though I would, certainly, prefer to
do it
all in one box. However, if I have to run power up the mast and stick
the
boat unit (vs the shore unit) someplace belowdecks, so long as I make
them
both happen off the same switch, it's of little consequence. I've got
an
instrument panel I've built to hide all the wires for everything at the
nav;
my 802 brain is behind the fixed part, and I could easily put the extra
unit
under the nav shelf behind my seat or in that part of the panel which
drops
down but currently is unoccupied (about half the space, leaving room
for
more instruments later).

So, I can live with that. Less (very little! - the circuit board and
duck
of the second unit doesn't weigh much!) weight aloft would be better,
along
with resolving any RFI issues of the two of them being proximate, too.
The
easiest way for me to reduce weight aloft would be to not use the cast
aluminum case he sent, instead some UV impervious plastic with a real
seal
rather than the pitiful gasket under the flat plate cover of the
aluminum
one.

>
> Now, combine all the above with trying to use other devices besides the
> WRT54g and, well, yeah it's going to be a challenge. It's all getting
> 'less
> worse' but it's still a far cry from being easy.

Meaning, the WRT54G is the only device which will work? I certainly
see an
awful lot of discussion about those units, so there's no question
there's an
awful lot of them out there...

Thanks for your input, and any further ideas you may have about solving
my
particular challenges. My time is running out, and getting the boat
floated
and seatrialed is higher on the list than this, but if I think it can
work,
I need to do it while I have the electrical and mast stuff opened up (I
have
to replace a fair amount of gear, and install other new stuff, up the
mast
once Lydia gets here at the end of July).

L8R

Skip

Morgan 461 #2
SV Flying Pig KI4MPC
http://tinyurl.com/p7rb4 - NOTE:new URL! The vessel as Tehamana, as we
bought her

The Society for the Preservation of Tithesis commends your ebriated
and scrutible use of delible and defatigable, which are gainly, sipid
and couth. We are gruntled and consolate that you have the ertia and
eptitude to choose such putably pensible tithesis, which we parage.

>>Stamp out Sesquipedalianism<<

Skip,

You cannot just choose random IP subnets. There are specific ranges you
should be using. The 15.15.15.0 subnet is a 'live' one. Use 'whois
15.15.15.0' and you'll see it's assigned to HP.

In Class C it's 192.168.x.x with a 255.255.255.0 mask. In Class B it's
172.16.x.x with a 255.255.0.0 netmask. Class A is 10.x.x.x with a 255.0.0.0
mask.

Some routing devices act to specifically handle these subnets differently.
Mainly to avoid bothering to publish them across WAN links and such.

Do not just choose something at random as there may be other things 'at
work' in how the routers handle things. It's VERY BAD IDEA to just go
inventing your own IP ranges. DO NOT DO IT.

> so long as I make them
> both happen off the same switch, it's of little consequence.

You may not want to have them on the same switch. When things like DHCP are
involved you generally want to make sure the issuing devices are up and
running BEFORE you bring up the others. Putting them all on the same
electrical switch might be a problem.

> Meaning, the WRT54G is the only device which will work?

Only device? No, but it, along with dd-wrt firmware, is one that's known to
be reliably hackable.

Honestly Skip from all the thread posts it really sounds like you're in over
your head. There just seems like you've been flailing around with too many
configuration options without really pinning down what does or doesn't work.
I could be wrong but I've been doing this a LONG time and have seen people
get themselves confused before...

Backtrack and get some reliable points of reference in what can or can't be
configured.

Up the mast you'll want a device configured to act as a 'client' of a shore
WiFi SSID. Down in the boat you'll want a device wired by ethernet to the
mast device, configured as an access point. The mast device should be
assigning DHCP, the in-boat device should not. The in-boat device should be
on a static IP address, one within the same range as the up-mast device.
Something like 192.168.202.0/255.255.255.0, for example, is "less likely" to
conflict with whatever IP ranges they might be using on the shore. You
could also use 172.16.202.0/255.255.0.0 if you wanted to be even less likely
to overlap with a shore subnet range. Note I'm using '202' but any number
in the 1-254 range could be used. The in-boat device should be setup on it's
own WiFi RF channel, generally different than the one up the mast. Nothing
should be acting as a "bridge".

Get these two things working. If the devices can't be configured to perform
these roles then you need to replace them with ones that can. Start simple,
build from there. Don't try to do everything at once.

What you'll end up with here is a 'double NAT' setup. You'll be publishing
DHCP addresses from the on-mast device through the in-boat access point,
translating them to the shore link, which will also be translating them
again out to their ISP. This may make it difficult, if not impossible, for
services beyond HTTP and mail to work. Stuff like Skype, instant messaging
or other multimedia services may not take kindly to being double-NAT'd. You
may have to forgo using those services. There may be more configuration
choices that could be used to get them working. But let's focus on getting
ANYTHING working FIRST, before getting into the more complex routing stuff.

Now, if you're insulted by the tone I'm taking here, don't be. The point
isn't to make you look or feel bad. The point is to cut through the
confusion and get the eff'ing crap working.

-Bill Kearney

Hi, Bill, and thanks for the followup,

Bill Kearney wrote:
> Skip,
>
> You cannot just choose random IP subnets. There are specific ranges you
> should be using. The 15.15.15.0 subnet is a 'live' one. Use 'whois
> 15.15.15.0' and you'll see it's assigned to HP.

Ah, well. In the early days of my attempting to make this work with
the vendor, internetzones.net, he had me doing all the flaky addresses
in an attempt to avoid the IP conflicts. I'd a whole lot rather use
the 192 sets...

>
> In Class C it's 192.168.x.x with a 255.255.255.0 mask. In Class B it's
> 172.16.x.x with a 255.255.0.0 netmask. Class A is 10.x.x.x with a 255.0.0.0
> mask.
>
> Some routing devices act to specifically handle these subnets differently.
> Mainly to avoid bothering to publish them across WAN links and such.
>
> Do not just choose something at random as there may be other things 'at
> work' in how the routers handle things. It's VERY BAD IDEA to just go
> inventing your own IP ranges. DO NOT DO IT.

Point taken. I dont' think I'll need to - as seen here and elsewhere,
I can't even get them to work individually, in a hardwired
configuration. Doing everything exactly plain-jane, factory default, I
can interrogate (for a while; not reliably, consistently, for an
extended period of time, such as a couple of hours), change the various
settings via URL interface, but not pass data (and not reliably
associate with any point, either - but when I *do* it won't pass data).

> > so long as I make them
> > both happen off the same switch, it's of little consequence.
>
> You may not want to have them on the same switch. When things like DHCP are
> involved you generally want to make sure the issuing devices are up and
> running BEFORE you bring up the others. Putting them all on the same
> electrical switch might be a problem.

Small enough issue, I suppose. Of course, it appears I'll have to
start over as to gear - I see in another thread that Senao distributors
are unlikely to accept returns; certainly, it's taking a major effort
to get mine returned, and only by getting my credit card comapany
involved.

>
> > Meaning, the WRT54G is the only device which will work?
>
> Only device? No, but it, along with dd-wrt firmware, is one that's known to
> be reliably hackable.

Heh. Close enough.

>
> Honestly Skip from all the thread posts it really sounds like you're in over
> your head. There just seems like you've been flailing around with too many
> configuration options without really pinning down what does or doesn't work.
> I could be wrong but I've been doing this a LONG time and have seen people
> get themselves confused before...

That's entirely possible. I am reasonably sure I've tried every single
possible configuration not less than a dozen times, including asking
them to behave in odd IP configurations as directed by the vendor who
likewise had no solutions to making these units work.

>
> Backtrack and get some reliable points of reference in what can or can't be
> configured.

I think I may have done that in the thread about "up the mast" - which
(making allowances for the likelihood that I'm in over my head) *seem*
to indicate the basic inoperability of the units as originally
designed, let alone interface to each other.

>
> Up the mast you'll want a device configured to act as a 'client' of a shore
> WiFi SSID. Down in the boat you'll want a device wired by ethernet to the
> mast device, configured as an access point. The mast device should be
> assigning DHCP, the in-boat device should not. The in-boat device should be
> on a static IP address, one within the same range as the up-mast device.

Clarification, please: I think I'm expecting the shore point to assign
an address. Is that done via *its* DHCP or the client bridge? If the
former, then, the bridge further assigns an address in *its* system to
the AP connected to it? And, assuming the up-the-mast is, for example,
192.168.1.100, then the AP (in the boat) would be something in the
family of 192.168.1.>100<255?

> Something like 192.168.202.0/255.255.255.0, for example, is "less likely" to
> conflict with whatever IP ranges they might be using on the shore. You
> could also use 172.16.202.0/255.255.0.0 if you wanted to be even less likely
> to overlap with a shore subnet range. Note I'm using '202' but any number
> in the 1-254 range could be used. The in-boat device should be setup on it's
> own WiFi RF channel, generally different than the one up the mast. Nothing
> should be acting as a "bridge".

Oops. No bridge on the mast? That's the point at which literally
everyone else has had me start (vendors, commentators here and
elsewhere, correspondents elsewhere). Understood about the 202 - the
vendor and I also did the 2xx groups trying to make it happen (again in
the early days - he quit talking to me after repeated failures and
repeated requests for return aurhorization).

And, while not a factor recently, assigning different channels was one
of the attempted cures (changing only one setting at a/each time and
testing again).

>
> Get these two things working. If the devices can't be configured to perform
> these roles then you need to replace them with ones that can. Start simple,
> build from there. Don't try to do everything at once.

If you've seen other recent posts about the same challenge, would you
say I had or had not defined that what I have won't work? I'll start
over if what I *can't* work. Certainly, that seems to be the case.

>
> What you'll end up with here is a 'double NAT' setup. You'll be publishing
> DHCP addresses from the on-mast device through the in-boat access point,
> translating them to the shore link, which will also be translating them
> again out to their ISP. This may make it difficult, if not impossible, for
> services beyond HTTP and mail to work. Stuff like Skype, instant messaging
> or other multimedia services may not take kindly to being double-NAT'd. You
> may have to forgo using those services. There may be more configuration

Aaack. That's half or more the reason I'm trying to make this work...

> choices that could be used to get them working. But let's focus on getting
> ANYTHING working FIRST, before getting into the more complex routing stuff.

Heh. LOL, actually. Agreed. Just getting some
raggasnagglegigafratzing wired bridge to work would be an achievement
at this point!

>
> Now, if you're insulted by the tone I'm taking here, don't be. The point
> isn't to make you look or feel bad. The point is to cut through the
> confusion and get the eff'ing crap working.

I'm not - and don't. And I'm entirely in agreement about the last
sentence.

So, unless you can see something I've missed, including the last couple
of posts in the "up the mast" thread where I do a step-by-step
instruction provided by John Navas, I assume that these Senao units are
good landfill fodder or (maybe, in their cards-out-of-the-board
configuration) pcmcia cards but nothing else useful to my objective?
And I should eat my losses and start over?

Thanks for bearing with me. I'm not really stupid - just stubborn. I
hate to give up on something which is *supposed* to work.

L8R

Skip
>
> -Bill Kearney

> Clarification, please: I think I'm expecting the shore point to assign
> an address. Is that done via *its* DHCP or the client bridge? If the
> former, then, the bridge further assigns an address in *its* system to
> the AP connected to it? And, assuming the up-the-mast is, for example,
> 192.168.1.100, then the AP (in the boat) would be something in the
> family of 192.168.1.>100<255?

Be careful about using the term 'bridge', it means different things. Few of
the WiFi devices actually behave reliably as a true bridge.

What I'm planning on doing is having the shore-link device get a single IP
address from the shore SSID. The shore-link device (up mast if you like)
will then act as a gateway to the on-boat device(s). One being an access
point wired to the shore-link device. The shore-link device will be handing
out DHCP addresses for it's own subnet.

192.168.12.0 on the shore SSID
172.16.12.0 on the mast-device
on-boat access point with it's own SSID
on-boat wired and wireless devices getting their address from the
mast-device

This means something on the boat gets an address from the boat's subnet and
is then routed to the shore link, which has it's own subnet and is then
routed out to the internet. If you use a Class B 172.16.0.0 subnet you're
pretty unlikely to ever run into overlap with the on-shore networks.

> Oops. No bridge on the mast? That's the point at which literally
> everyone else has had me start (vendors, commentators here and
> elsewhere, correspondents elsewhere). Understood about the 202 - the
> vendor and I also did the 2xx groups trying to make it happen (again in
> the early days - he quit talking to me after repeated failures and
> repeated requests for return aurhorization).

Yes, well, 'bridging' to the shore subnet may not be as reliable as you'd
like. If you use an on-boat subnet then you never have to reconfigure those
devices. Yes, you end up being behind a second NAT router but most
protocols handle this without too much trouble. This is also where things
like WDS are "supposed to" be able to work. But they run afoul of too many
variables. When you're talking about bathing a floor of cubicles in WiFi
signals stuff like WDS can be made to work. But when you're talking about
an unstable platform like a boat, all bets are off. You're faced with
signal reliability at the start. Then you run into the problem of not
controlling both sides of the networking equation. Far too often you're
going to run into a shore device that doesn't have the right firmware or
setup to actually make things work. And by the time you figure this out,
well, lets just say I'd rather not waste that much time/effort.

I may even go so far as to tunnel from my on-boat subnet back to my office
via VPN. The dd-wrt firmware supports doing this on the router itself.
That way I'm only using the shore link as a tunnel to push all my packets
back to the office where I've got control over the uplink back out to the
internet. Sure, there's a performance hit to do this but for some of the
stuff I'd be doing it's not a big deal. Just something to think about.
Remember, packets over clear air are sniffable. If you tunnel them through
an encrypted VPN you eliminate a considerable amount of risk.

And if you bridge your on-boat devices as direct members of the on-shore
subnet then you're at the mercy of whatever security, or lack thereof, of
that network. When you setup a NAT router on the boat you're a bit safer
being behind it's firewall. This is also why you may not want to 'share'
your on-boat network with others. While it's a nice idea not everyone else
on those boats may be as altruistic as you'd like. Regardless, always run a
software firewall on your computers anyway. Don't assume anything's ever
safe enough.

> And, while not a factor recently, assigning different channels was one
> of the attempted cures (changing only one setting at a/each time and
> testing again).

Also configure the on-boat WiFi to use as weak a signal as necessary. While
it's laudable to think about bathing the entire anchorage in a shore-link
you're also asking for difficulties in just getting your own stuff working.
I'm all for being generous, but I'd want things to actually WORK first.

> If you've seen other recent posts about the same challenge, would you
> say I had or had not defined that what I have won't work? I'll start
> over if what I *can't* work. Certainly, that seems to be the case.

Frankly, and this is rude to say, but your posts have been quite difficult
to figure out. It seemed you were flailing about with too many
configuration variables.

So no, I don't have any real idea what you've defined. And I don't want to
either. Better to 'punt' and start over from the basics. Get the devices
working to within their known configurations and THEN start mucking it up.

> Aaack. That's half or more the reason I'm trying to make this work...

And you've got nothing working now.

> Heh. LOL, actually. Agreed. Just getting some
> raggasnagglegigafratzing wired bridge to work would be an achievement
> at this point!
>
> > Now, if you're insulted by the tone I'm taking here, don't be. The
point
> > isn't to make you look or feel bad. The point is to cut through the
> > confusion and get the eff'ing crap working.
>
> I'm not - and don't. And I'm entirely in agreement about the last
> sentence.

Good, some folks can't grasp that text doesn't always convey sentiments
accurately.

> Thanks for bearing with me. I'm not really stupid - just stubborn. I
> hate to give up on something which is *supposed* to work.

Well, technological devices aren't always broken when people think they
'ought to be able' to do something. And support personnel don't always have
a clue about anyhing beyond the most basic of configurations. Combine the
two and you've got a recipe for frustration.... but you know this already.

So take a step back and get some fundamentals figured out. Get some of the
boxes you've got working in their basic configurations. Make sure they can
actually do the basic things they're supposed to provide. Only then should
you start mucking around further with them. I suspect the device you've got
can be configured to make some of this work. Get them working at a basic
level first and then see what else needs to be added, configured or
replaced.

-Bill Kearney