Re: [pfSense Support] OT: Re: Tricky infrastructure question

2008-12-01 Thread Glenn Kelley

Depending on budget I can mention a number of alternatives.

Check out the Motorola Canopy - or perhaps a Tranzeo unit.

Low cost - and you can push 110mb plus over these units depending on  
what one you get.


I like the Moto units because they are not the normal 802.11..x stuff


On Dec 1, 2008, at 2:52 AM, Chris Bagnall wrote:

We are currently using vlans because we have VoIP services going  
through
this and different kind of users.  Everything is working OK as of  
now.
However, the max bandwidth of one WiFi link like that is about 10  
mbps.

To increase the total bandwidth, we want to add another antenna in
Building 1.


I don't know the layout of the buildings, but have you considered  
using point-to-point laser links between the buildings? I was  
involved in a project at a site earlier this year that was happily  
getting 100mbps between buildings using roof-mounted lasers.


Much lower latency than 802.11a/b/g as well (assuming that's what  
you're using).


If you do stick with WiFi, I'd not worry too much about the downtime  
- a) it should only be a few minutes, and b) it can probably be  
managed such that it's done during off peak hours.


Can you give us any more information about the WiFi hardware at each  
end? I'm not sure simple port trunking on the switches (which is  
what I guess you're suggesting in the first approach) is going to do  
the job if the WiFi devices are essentially standard APs, since I  
presume they'd have their own IP addresses?


If the WiFi devices were operating at a lower level than IP, it'd  
probably work, possibly if they were using some sort of L2TP  
tunnelling internally? (though you might then have to consider  
packet fragmentation, depending on the packet size)


Regards,

Chris



-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org




-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



[pfSense Support] Re: OT: Re: Tricky infrastructure question

2008-12-01 Thread Ugo Bellavance

Michel Servaes a écrit :
I can only think of using a switch, being capable of port bonding... 
802.3ad capable switches like HP Procurve 1800's can link multiple ports 
for better speed.

Don't know how they end up, using wireless bridges though ;-)


Thanks for your input, but the last line is also my concern :).


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



[pfSense Support] Re: OT: Re: Tricky infrastructure question

2008-12-01 Thread Ugo Bellavance

Chris Bagnall a écrit :

We are currently using vlans because we have VoIP services going through
this and different kind of users.  Everything is working OK as of now.
However, the max bandwidth of one WiFi link like that is about 10 mbps.
 To increase the total bandwidth, we want to add another antenna in
Building 1.


I don't know the layout of the buildings, but have you considered using 
point-to-point laser links between the buildings? I was involved in a project 
at a site earlier this year that was happily getting 100mbps between buildings 
using roof-mounted lasers.

Much lower latency than 802.11a/b/g as well (assuming that's what you're using).

If you do stick with WiFi, I'd not worry too much about the downtime - a) it 
should only be a few minutes, and b) it can probably be managed such that it's 
done during off peak hours.

Can you give us any more information about the WiFi hardware at each end? I'm 
not sure simple port trunking on the switches (which is what I guess you're 
suggesting in the first approach) is going to do the job if the WiFi devices 
are essentially standard APs, since I presume they'd have their own IP 
addresses?

If the WiFi devices were operating at a lower level than IP, it'd probably 
work, possibly if they were using some sort of L2TP tunnelling internally? 
(though you might then have to consider packet fragmentation, depending on the 
packet size)


The antennas are SkyPilots.  They do have an IP address, for 
configuration, but I think they could also be configured via console.


Thanks a lot for your input on this OT matter :).

Ugo


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



[pfSense Support] Monitor IP address

2008-12-01 Thread Mike Lever
Hi,

Can somebody please explain to me exactly how this works. I am having an
argument with my superior. He is insistent on setting the monitor IP
addresses in my load balancer pool to the same IP address. In his mind it
makes sense, as that way it will pick up which line is the fastest to the
same point and route accordingly. 

I read in the manuals that these IP addresses should be unique, and
therefore did as the manual said. What will happen if they are set to the
same address and why is that so ? 

Here is my thinking on how it works, please correct me where I am going
wrong. 

I have 5 WAN ports. The load balancer will constantly ping WAN1, WAN2,WAN3,
WAN4  WAN5 simultaneously. Depending on which has the quickest response and
is not currently transmitting packets, it will utilise. Then why set the
unique IP addresses ?

Best regards,
Mike



Mike Lever




+27 82 903 8613  –  Mobile
+27 11 807 0100  –  Telephone
+27 11 807 1208  –  Fax


http://www.velocityfilms.com 
  


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



Re: [pfSense Support] Monitor IP address

2008-12-01 Thread Bill Marquette
On Mon, Dec 1, 2008 at 2:41 PM, Mike Lever [EMAIL PROTECTED] wrote:
 Hi,

 Can somebody please explain to me exactly how this works. I am having an
 argument with my superior. He is insistent on setting the monitor IP
 addresses in my load balancer pool to the same IP address. In his mind it
 makes sense, as that way it will pick up which line is the fastest to the
 same point and route accordingly.

Yeah, that won't work.

 I read in the manuals that these IP addresses should be unique, and
 therefore did as the manual said. What will happen if they are set to the
 same address and why is that so ?

You'll actually lose link failure detection.  Whichever link came up
last will set the route to your monitor IP through it.

 Here is my thinking on how it works, please correct me where I am going
 wrong.

 I have 5 WAN ports. The load balancer will constantly ping WAN1, WAN2,WAN3,
 WAN4  WAN5 simultaneously. Depending on which has the quickest response and
 is not currently transmitting packets, it will utilise. Then why set the
 unique IP addresses ?

Usually the monitor IP is set to the next hop so you can detect link
failure.  Latency is not taken into account.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



RE: [pfSense Support] Monitor IP address

2008-12-01 Thread Mike Lever
Thanks for the explanation Bill. 

Can you please elaborate where you mention: 

You'll actually lose link failure detection

What exactly is link failure detection ? I understand the meaning of the
words in isolation but can you elaborate in the load balancing / Pfsense
context ? 

Whichever link came up last will set the route to your monitor IP through
it.

So then, say WAN2 was the last WAN port to come up and the monitor addresses
were set to the same IP address, would it then only route traffic through
WAN2 ? 

Best regards,
Mike



Mike Lever




+27 82 903 8613  -  Mobile
+27 11 807 0100  -  Telephone
+27 11 807 1208  -  Fax


http://www.velocityfilms.com 

-Original Message-
From: Bill Marquette [mailto:[EMAIL PROTECTED] 
Sent: 01 Dec 2008 10:46 PM
To: support@pfsense.com
Subject: Re: [pfSense Support] Monitor IP address

On Mon, Dec 1, 2008 at 2:41 PM, Mike Lever [EMAIL PROTECTED] wrote:
 Hi,

 Can somebody please explain to me exactly how this works. I am having an
 argument with my superior. He is insistent on setting the monitor IP
 addresses in my load balancer pool to the same IP address. In his mind it
 makes sense, as that way it will pick up which line is the fastest to the
 same point and route accordingly.

Yeah, that won't work.

 I read in the manuals that these IP addresses should be unique, and
 therefore did as the manual said. What will happen if they are set to the
 same address and why is that so ?

You'll actually lose link failure detection.  Whichever link came up
last will set the route to your monitor IP through it.

 Here is my thinking on how it works, please correct me where I am going
 wrong.

 I have 5 WAN ports. The load balancer will constantly ping WAN1,
WAN2,WAN3,
 WAN4  WAN5 simultaneously. Depending on which has the quickest response
and
 is not currently transmitting packets, it will utilise. Then why set the
 unique IP addresses ?

Usually the monitor IP is set to the next hop so you can detect link
failure.  Latency is not taken into account.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



Re: [pfSense Support] Monitor IP address

2008-12-01 Thread Chris Buechler
On Mon, Dec 1, 2008 at 3:41 PM, Mike Lever [EMAIL PROTECTED] wrote:

 I have 5 WAN ports. The load balancer will constantly ping WAN1, WAN2,WAN3,
 WAN4  WAN5 simultaneously. Depending on which has the quickest response and
 is not currently transmitting packets, it will utilise.

What Bill said is correct. One additional comment, the above isn't
true. Your load balancing is round robin, all connections in a pool
are used equally. If the monitor IP for a specific gateway stops
responding, that gateway is removed from the pool. If you use the same
monitor IP for all connections, it won't work as when that monitor IP
stops responding, the system will think you have no WANs available and
remove them all.

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



Re: [pfSense Support] Monitor IP address

2008-12-01 Thread Bill Marquette
On Mon, Dec 1, 2008 at 3:06 PM, Mike Lever [EMAIL PROTECTED] wrote:
 Thanks for the explanation Bill.

 Can you please elaborate where you mention:

 You'll actually lose link failure detection

 What exactly is link failure detection ? I understand the meaning of the
 words in isolation but can you elaborate in the load balancing / Pfsense
 context ?

Only one of the links (whichever one has decided that your monitor IP
is available over it) will actually do any link failure detection.
ie.  in your case with 5 WANS, if monitoring is occurring for WAN5 and
it's the same address as WANS1-4, if WAN1 goes down, you'll still send
1/5th of your traffic down that pipe (even though it won't work) as
there will be nothing in place to determine it's availability.

 Whichever link came up last will set the route to your monitor IP through
 it.

 So then, say WAN2 was the last WAN port to come up and the monitor addresses
 were set to the same IP address, would it then only route traffic through
 WAN2 ?

It'll still round robin over all 5 links.  It's just that only one of
them will be monitored for availability.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



Re: [pfSense Support] Monitor IP address

2008-12-01 Thread Bill Marquette
On Mon, Dec 1, 2008 at 3:09 PM, Chris Buechler [EMAIL PROTECTED] wrote:
 On Mon, Dec 1, 2008 at 3:41 PM, Mike Lever [EMAIL PROTECTED] wrote:

 I have 5 WAN ports. The load balancer will constantly ping WAN1, WAN2,WAN3,
 WAN4  WAN5 simultaneously. Depending on which has the quickest response and
 is not currently transmitting packets, it will utilise.

 What Bill said is correct. One additional comment, the above isn't
 true. Your load balancing is round robin, all connections in a pool
 are used equally. If the monitor IP for a specific gateway stops

This is an important point to note.  Monitoring is for the purposes of
availability, not for latency detection.  The WANs are load balanced
from a connection perspective, not from a throughput or latency
perspective.  If you have a single flow eating up an entire
connection, nothing will stop other flows from using that connection.
The load balancing is on a flow by flow basis in a round robin
fashion.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



[pfSense Support] RE: [Pfsense Support] Monitor IP address

2008-12-01 Thread Mike Lever
Great, thank you very much Bill. 

One point for clarification purposes... please define a flow ? 

Best regards,
Mike



Mike Lever




+27 82 903 8613  -  Mobile
+27 11 807 0100  -  Telephone
+27 11 807 1208  -  Fax


http://www.velocityfilms.com 

-Original Message-
From: Bill Marquette [mailto:[EMAIL PROTECTED] 
Sent: 02 Dec 2008 12:33 AM
To: support@pfsense.com
Subject: Re: [pfSense Support] Monitor IP address

On Mon, Dec 1, 2008 at 3:09 PM, Chris Buechler [EMAIL PROTECTED] wrote:
 On Mon, Dec 1, 2008 at 3:41 PM, Mike Lever [EMAIL PROTECTED]
wrote:

 I have 5 WAN ports. The load balancer will constantly ping WAN1,
WAN2,WAN3,
 WAN4  WAN5 simultaneously. Depending on which has the quickest response
and
 is not currently transmitting packets, it will utilise.

 What Bill said is correct. One additional comment, the above isn't
 true. Your load balancing is round robin, all connections in a pool
 are used equally. If the monitor IP for a specific gateway stops

This is an important point to note.  Monitoring is for the purposes of
availability, not for latency detection.  The WANs are load balanced
from a connection perspective, not from a throughput or latency
perspective.  If you have a single flow eating up an entire
connection, nothing will stop other flows from using that connection.
The load balancing is on a flow by flow basis in a round robin
fashion.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



Re: [pfSense Support] RE: [Pfsense Support] Monitor IP address

2008-12-01 Thread Bill Marquette
On Mon, Dec 1, 2008 at 4:42 PM, Mike Lever [EMAIL PROTECTED] wrote:
 Great, thank you very much Bill.

 One point for clarification purposes... please define a flow ?

Any given TCP connection (from connection setup, to teardown).  Or UDP
(say a VOIP call) stream of sufficient packet frequency to remain in
state.

--Bill

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Commercial support available - https://portal.pfsense.org



[pfSense Support] IPSEC stopped working

2008-12-01 Thread Paul
IPSEC with shrewsoft has been working great and all of a sudden I cant 
bring the tunnel up with the following log in pfsense.Nothing has 
changed that I could point to


racoon: *[Self]*: INFO: 66.x.x.x[500] used as isakmp port (fd=20)
Dec 2 00:24:55 	racoon: INFO: fe80::207:e9ff:fe07:c085%fxp0[500] used as 
isakmp port (fd=19)
Dec 2 00:24:55 	racoon: *[Self]*: INFO: 192.168.2.1[500] used as isakmp 
port (fd=18)
Dec 2 00:24:55 	racoon: INFO: fe80::21b:21ff:fe1b:4c92%em0[500] used as 
isakmp port (fd=17)
Dec 2 00:24:55 	racoon: *[Self]*: INFO: 127.0.0.1[500] used as isakmp 
port (fd=16)





Thank you for any help

paul