That is wise. None of ours NAT, but at an apartment complex we made an
outdoor FlashROM as a satellite station (works great as far as the
elements are concerned) and received a wireless backhaul which fed the
complex tenants via Ethernet. The key factor was that we assumed the
FlashROM could handle NAT for 150 clients. It could/did, until the NAT
translations pegged. Fortunately we learned that our outdoor FlashROM
works through the Texas heat, as well as tropical storm rains.
Steve Deaton
IT Director
Texas Broadband, Inc.
(888)868.3835 ext 85 (office)
979.289.0148 (office)
979.289.5117 (fax)
-----Original Message-----
From: karlnet-bounces@WISPNotes.com
[mailto:karlnet-bounces@WISPNotes.com] On Behalf Of Bob Hrbek
Sent: Wednesday, September 10, 2003 2:26 PM
To: Karlnet Mailing List
Subject: Re: [Karlnet] Strange ping loss
most of my AP's either route or are just dumb bridges. I don't think
they
are powerful enough to handle the NAT.
----- Original Message -----
From: "Steve Deaton" <steve@texasbb.com>
To: "'Norm Young'" <npyoung@applegatebroadband.net>; "'Karlnet Mailing
List'" <karlnet@WISPNotes.com>
Sent: Wednesday, September 10, 2003 2:20 PM
Subject: RE: [Karlnet] Strange ping loss
> For whatever it's worth, I just found a NAT problem. This will only
> apply if you're running NAT on the device that is locking up. Some
> clients behind the NAT AP were running something that was generating a
> large number of NAT connections. I wasn't able to see this until I
> bridged the AP to a Cisco router. I looked at the number of NAT
> translations and there were tens of thousands. The AP running NAT was
> not able to keep up with the large number of NAT translations so it
> locked. When bridged, my 7206 would run at 98% CPU utilization. I
shut
> off the MAC addresses that were generating the most connections and
all
> was well. An alternate solution is to give everyone a public, thus
> negating the NAT and all is well now too. This maybe something that
> everyone already knows, but it was a simple thing that I overlooked,
so
> maybe others have also. Hope this helps.
>
> Steve Deaton
> IT Director
> Texas Broadband, Inc.
> (888)868.3835 ext 85 (office)
> 979.289.0148 (office)
> 979.289.5117 (fax)
>
>
> -----Original Message-----
> From: karlnet-bounces@WISPNotes.com
> [mailto:karlnet-bounces@WISPNotes.com] On Behalf Of Norm Young
> Sent: Wednesday, September 10, 2003 10:25 AM
> To: Karlnet Mailing List
> Subject: Re: [Karlnet] Strange ping loss
>
> Hi Bob,
>
> I think that I may be seeing the same thing. I noticed that sometime
> this
> summer as I added clients to our WIPOP the jitter from the clients to
> the
> AP/NOC really went bad (as previously mentioned) AND I'm seeing
timeouts
> on
> my SNMP traffic grapher occasionally, and if I do a ping from the
> clients to
> the NOC. I'm going to try and replace the AP1000 (relegating it to
> backhaul) with a WPG01 (and cross my fingers and hope I don't have
> lockup
> problems), while splitting the clients across two radios/antennas and
> see
> what that does. The amount of loading (according to what Karlnet
> states is
> possible) on this base is moderate, say 20 clients! I've already
tried
> updating firmware, trying new radios at the base, all without any
> change.
>
> Norm
>
> ----- Original Message -----
> From: "Bob Hrbek" <bhrbek@jagwireless.net>
> To: "Karlnet Mailing List" <karlnet@WISPNotes.com>
> Sent: Wednesday, September 10, 2003 6:43 AM
> Subject: [Karlnet] Strange ping loss
>
>
> > I have an Orinoco COR v4.15 with about 14 RG-1000 clients and 1
> AP-1000
> that
> > is serving as a POP and backhaul for the small pop. In all, about
35
> > clients are going through this AP, in on way or another.
> >
> > I have a monitoring system that pings all the clients every 2
minutes.
> > (Because of this problem, I stuck a monitoring box at the bottom of
> this
> > tower) It keeps reporting the clients as un-reachable. I've not
> contacted
> > any of them but had one complaint off the remote AP-1000
> Satellite/Base.
> >
> > My testing:
> >
> > I did a 1000 byte ping to one client. I left it running all night.
> >
> > --- 10.100.105.16 ping statistics ---
> > 43477 packets transmitted, 43474 received, 0% loss, time 43582095ms
> > rtt min/avg/max/mdev = 9.543/15.222/311.773/11.513 ms
> >
> >
> > Out of 43477, I lost 3 packets!!!!! My max ping time was 311ms.
The
> client
> > I was pinging was the only one that I never received an outage
> notification
> > on last night. Previously I was getting 20 or more per day.
> >
> > The stats on the AP don't look wonderful. Re-transmits are about
10%
> for
> > the clients that I get the most outages reports on. Overall in FCS
> errors
> > are only .3% (yes that is point 3%) of the total inbound packets. I
> believe
> > I have some RF issues at this site but WHY, if I ping a client
> constantly,
> > will I be successful 43,374 times out of 43,477 and still drop so
many
> > monitoring pings when I'm not constantly pinging him? I even
disabled
> > polling on this ap and it didn't change anything but lower the ping
> times.
> >
> > If I didn't know better, I'd say I had a duplicate IP address but it
> is
> > affecting about 10 of the RG"s on a regular basis. No dupes out
> there. I
> > checked.
> >
> > Any thoughts?
> >
> > -Bob
> >
> > _______________________________________________
> > Karlnet mailing list
> > Karlnet@WISPNotes.com
> > http://lists.wispnotes.com/mailman/listinfo/karlnet
> >
>
>
> _______________________________________________
> Karlnet mailing list
> Karlnet@WISPNotes.com
> http://lists.wispnotes.com/mailman/listinfo/karlnet
>
> _______________________________________________
> Karlnet mailing list
> Karlnet@WISPNotes.com
> http://lists.wispnotes.com/mailman/listinfo/karlnet
>
_______________________________________________
Karlnet mailing list
Karlnet@WISPNotes.com
http://lists.wispnotes.com/mailman/listinfo/karlnet
|