our solution involves a Cisco router instead.
Also, if your co-worker is getting drop-outs on calls, he probably has
other issues. I've held an hour-long conversation with no hiccups before
while downloading at 28 Mbit/sec the entire time.
George Metz
Ken Gentle wrote:
> Thanks, Charles.
Andrew Haninger wrote:
> On 2/26/07, George Metz <[EMAIL PROTECTED]> wrote:
>> Any number of things could cause that. My concern is, sometimes
>> Motorolas don't handle forced duplex settings well, and sometimes
>> half-duplex will make things problematic to
bably
not the hub, or if it is then it's probably just the port itself, so try
a different port too.
George Metz
-
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and
Andrew Haninger wrote:
> I half understand duplex (don't bother explaining it - I'll look it up
> for myself) but I don't understand how duplex could be negotiated
> improperly?
Any number of things could cause that. My concern is, sometimes
Motorolas don't handle forced duplex settings well, and
Andrew Haninger wrote:
> I don't think I explained the "different IP" well enough. I get
> assigned an from a completely different range IP and also a different
> gateway:
>
> Slow LEAF box:
> 71.72.x.x/22 dev eth0 proto kernel scope link src 71.72.x.x
> default via 71.72.96.1 dev eth0
>
I can confirm 100% that the 3c59x.o module works with the 3c905C 3Com
NIC. Said module and NIC are in use as I type this on eth0 on my Bering box.
George
[EMAIL PROTECTED] wrote:
Eric,
The modules tarball for the kernel 2.4.31 does not include 3c905c. It
includes 3c501, 503, 505, 507, 509, 51
orge
Arne Bernin wrote:
On Wed, 2005-09-14 at 23:59 -0400, George Metz wrote:
Hi George!
That's odd.
I actually used that page, but it didn't supply the crc32.o without
being manually told to use it. I doublechecked lsmod and tulip is the
only thing relying on it.
I suppose you u
That's odd.
I actually used that page, but it didn't supply the crc32.o without
being manually told to use it. I doublechecked lsmod and tulip is the
only thing relying on it.
Looks like, checking the modules.dep, that the change occurred between
2.4.20 from release 2.0 and 2.4.26 from 2.2.0
Might I make a suggestion here?
There should be some form of documentation, either in the installation
doc or on the /etc/modules file, stating that tulip.o for Bering uClibC
has dependencies within crc32.o. I spent a profitable couple of hours
banging my head on that issue with 2.3 rc1, event
Thanks Mike, that would explain it.
George
Mike Noyes wrote:
On Sat, 2005-09-10 at 13:49, George Metz wrote:
Getting the following as a text line when I try and load either
www.leaf-project.org or leaf.sourceforge.net:
Unable to load database indicated by configuration file.
No errors
Getting the following as a text line when I try and load either
www.leaf-project.org or leaf.sourceforge.net:
Unable to load database indicated by configuration file.
No errors, just that one line.
Incidentally, whatever's going on, it's been going on long enough that
the 4th unique result fr
le video streams going through this router/firewall
nearly 24/7. (i.e. Lots of bandwidth, very few connections) Do you think I
need the extra cpu of a regular computer or will the wrap be able to handle
it?
Thanks,
Andrew
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTE
None of the over-the-counter router-in-a-boxes are going to be able to
handle multiple static IPs, with the possible exception of a Linksys
that's had it's firmware replaced with a Linux-based one from the
hardware hacking groups.
An entry level Cisco is hideously expensive; I found two on Pri
Your problem is that the 3c905C series cards use the 3c59x.o module, not
the 3c90x.o one. I've no idea why the architecture was changed that
drastically with only a single letter to mark the difference, but it
does, and I used a 905C as a main ethernet card for 3 years on the 3c59x
module so yo
Anyone know if there's an already-compiled module out there for Bering
or Bering uClibC for the Zoom Model 5001 PCI cablemodem? There's a
reference driver up on their website that's distributed under the GPL;
I'm just wondering if anyone's grabbed it.
I'm looking to buy a new cablemodem, and no
Chris Lee wrote:
As I don't know how to config CISCO, I use Getif to peek the config via
SNMP.
For Interface, it show:
descr. ip address
Ethernet0 10.0.108.254/255.255.255.0
203.198.77.78/255.255.240
Ethernet1 172.23.76.154/255.255.255.252
Tunnel5 192.16
Because not everyone uses Linux on anything but their LEAF box. Heck, I
DO use Linux, and when I do an upgrade it's usually with WinImage.
K.-P. Kirchdörfer wrote:
Am Dienstag, 11. Mai 2004 20:04 schrieb Marko Nurmenniemi:
K.-P. Kirchdörfer wrote:
Due to new linuxrc "backupdisk" is broken and h
Couple of things to let you know about:
Jim Hubbard wrote:
1. What sort of throughput, for instance, could LEAF-Bering theoretically
provide on a Pentium 100 system with edo ram and with 10/100 nics, cables,
and switch, assuming that all other systems connected have unlimited speed?
Check the arc
Don't know about shorewall (which you would have to configure to allow
VPN traffic to pass through to that specific IP address), but what you
basically want it to do is substitute for a traditional router.
Effectively, you'd simply have to turn off NAT and let DNS and the
public IP addresses do
This is not entirely correct. There is, in fact, an RFC1918 equivalent
for AS routing numbers, for one. Of course, a private AS should only
really be used if you're multihoming to two different gateways on the
same provider network.
Additionally, ARIN only requires that you have a unique routin
The first thing I'd be doing here is NOT asking how to allow these
packets to pass, but trying to figure out why they're being sent in the
first place.
If you're using a default Bering install without monkeying with the
Bering settings, and you're using DHCP, then your gateway should be
192.16
You shouldn't be, because you were right that there's a bottleneck, you
just missed what the bottleneck is.
A 768k T-1 or Cablemodem line is going to give you around
90-95Kbytes/sec on a download, whereas your DSL is only turning out
around 70Kbytes/sec. The reason for this is pretty straightfo
Really wouldn't matter, just yet, that ping isn't enabled. If he has no
link light from the LEAF box on his switch/hub, and no light on the
networking card, then there's an issue that's lower than Layer 3 (IP),
and probably an issue at Layer 1 (Physical).
After he figures out the reason he does
Step 1: Doublecheck your cable. Try swapping the cable on eth1 for the
cable on eth0, and vice versa, and see if the lights follow the cables.
If they do (eth0 dark, eth1 lit) then replace the bad cable.
That's the only thing that leaps to mind, probably because I had the
same problem with my o
Honestly, you should probably NOT be using the root servers. They're in
general designed to provide updates to other DNS servers on the net, and
in the case of on-network resources that may not have a publically
routable IP address, going to the root server is going to give you an IP
address th
es the way LEAF handles stuff compared to LRP.
George Metz wrote:
Looks like Issue 100 (August 2002) of Linux Journal. I'm having a bit of
difficulty actually pulling it up on their website, however. That might
be just a slow load... ah, there we go. Just give it a little bit and it
will com
Looks like Issue 100 (August 2002) of Linux Journal. I'm having a bit of
difficulty actually pulling it up on their website, however. That might
be just a slow load... ah, there we go. Just give it a little bit and it
will come up just fine.
And frankly, having worked with both LEAF (specifical
If this is a roleplaying game like Everquest, Asheron's Call, Dark Age
of Camelot, Horizons, etc. then most likely no configuration will need
to be done. Most of the massively-multiplayer RPGs out there work on a
single outbound UDP connection, with multiple inbound UDP connections in
response.
It depends on which of the three traffic light graphics is red. If it's
the Firewall light, this means there is probably a bunch of traffic that
your firewall has rejected.
If you want to see the traffic that is being dropped, click on the
"shorewall.log" link on the page that tells you there's
If I recall correctly, this was looked at in the past and the cost was
prohibitive, to say the least. A quick poke around on Google isn't
turning up an amount, but I remember it being significantly more than
was worth the effort.
Not to mention that ICSA Certification is designed for commercial
7;re sending to the WAP in the first place.
George
Mike Noyes wrote:
On Fri, 2003-12-19 at 10:16, Mike Noyes wrote:
On Fri, 2003-12-19 at 09:24, George Metz wrote:
Yeah, I know. I was more replying to someone else saying that WEP was
enough. It's clearly not.
George,
Is NoCatAuth/NoCat
tion is so flawed, and no infrastructure to
change keys when they have been compromised. That's why IPSEC is so
important.
Sean
On Thu, 2003-12-18 at 12:19, George Metz wrote:
The problem with this approach is that WEP, the security protocol that
most Wireless points use, is fairly weak and
The problem with this approach is that WEP, the security protocol that
most Wireless points use, is fairly weak and relatively easily broken.
If you want to ensure that only authorized users can get in, you kind of
want to use both WEP (Wired Equivalent Protocol, even though it's not...
:) ) an
Couple of things on this. Interspersed where relevant.
Brian Kolaci wrote:
Hi,
I'm looking to setup a box mainly as a routing decision maker.
I'll have 2 DSL lines, a primary and backup (to 2 different ISP's). I'd
like traffic to go out the primary (faster and static IP's) when its up
and have i
This cannot be a DNS issue. It's like saying, every time a plane flies
over my house, the subway train that runs underneath it gets derailed.
DSL modem sync is a Layer 2 function, whereas DNS is a Layer 7(?)
function. (I'm talking about the OSI Layer Model. Layer 2 is Data Link,
Layer 7 is appl
Negative, 192.168.0/23 will route 192.168.0.0/24 and 192.168.1.0/24
but ignore 192.168.2.0/24. You'd need to do a /22 to do aggregate
routing with the specified /24s, and at that you'd have two /24s
floating in limbo.
Victor McAllister wrote:
Dachstein will not route between interfaces unless y
RFC1483 is simply a means - that is as far as you're concerned,
completely and totally transparent - for encapsulating IP traffic for
transport over an ATM circuit. The DSL Modem handles all of this, and
it has absolutely zilch to do with your IP address or even with
anything that actually touc
Just as a note, my primary reasoning for thinking to put NAT behind
NAT - and it wouldn't be an issue, BTW, since many ISP/MSP/MSSP
companies, including the one I work for, provide RFC1918 address space
for the WAN side and run NAT behind it on the LAN side, because it's
all going out a managed
If you didn't compile the module with the kernel that Bering uses
running, or with links pointing to the Bering kernel's config and
source directories, you will not be able to successfully get the
module to work.
Sebastian A. Aresca wrote:
Hi i buy a D-Link 520+ and I am trying to make it work
ugh, in routing mode, you can't
play Asheron's Call - one of Microsoft's games - from more than one
client at a time. I'm assuming this would be an issue as a router for
any online games that use multiple UDP connections. Bering 1.2 and
Shorewall handle it out of the box, a
From experience on the far end of the line, a T-1 can hold anywhere
from 2 to 1000 users simultaneously. If this setup is mostly web
browsing for research purposes - for example, a school library - where
they either won't want to or won't be able to run their own programs
on it, then a frac T-1
- Original Message -
From: Ray Olszewski <[EMAIL PROTECTED]>
Date: Wednesday, August 6, 2003 10:13 am
Subject: Re: [leaf-user] Routing 192.168.231.255
> As an aside, from time to time people post questions here about
> whether
> blocked packets from/to ports 137/139 are attacks. I usuall
Um.
Okay, color me stupid.
For some reason, I mixed up pump and DHCPd for some reason. DHCPd is
not loaded in there by default, but is on the disk.
Sorry for confusing you...
George Metz
Luis.F.Correia wrote:
If you are using Bering, then type 'ip addr'.
-Original Message
To clarify, however...
Bering is indeed setup to use pump.lrp by default, and it works
extremely well. HOWEVER, since Bering is set up so that you can use
DHCP, PPP, or PPPoE with the default image, pump.lrp is NOT loaded by
default in syslinux.cfg.
So, if you open up syslinux.cfg and add pump
or you.
That's the real threat. Granted, maybe not the entire net, but a far
larger portion than you'd like to think is healthy.
George Metz
- Original Message -
From: Eric B Kiser <[EMAIL PROTECTED]>
Date: Tuesday, July 29, 2003 7:42 pm
Subject: RE: [leaf-user] VPN security
anicky. It's certainly a strong likelihood, and AFAIK there's
relatively little chance of the hole you're referring to from
happening. (IOW, browsing on your public connection while connected
via VPN.)
George Metz
---
This SF.Net em
Steve Wright wrote:
I know Linux and WISP-DIST, but I am not familiar with Bering, per se.
Try these things ;
boot Bering, and see if pump is running on eth0. It should not be.
Start from scratch, with a perfectly clean image and have another go.
Write a little script that that restarts the n
use their own chipsets have modules that start with "3c"
as a designator.
George Metz
---
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Downloa
As a start, I'd like to say, please take the politics off this list.
Matt did. You can too. Furthermore:
I was working for the french gov and I have very good access to some
non-public informations... The European had scanned by satelit the
Near-East (from Syria to Iran) for Radioactivity...
Michelle Konzack wrote:
No I have not...
because I curently no running SLINK-System (HD crash)
and can not build new LRP 2.9.4 Packages...
You could try the windows port of it, WinDump. It runs on
Win9x/ME/NT/2K/XP, and only requires a (freely available) single file
in addition to itself.
Jeff Newmiller wrote:
On Sun, 20 Jul 2003, George Metz wrote:
I'd like to do it on CD-ROM. And the easiest way for me to make the
infamously-missing initrd.cdrom file myself would be to extract initrd
on my workstation, make the changes, and close it back up.
Unfortunately, things se
image on floppy and borrowing a computer?
Thanks!
George Metz
---
This SF.net email is sponsored by: VM Ware
With VMware you can run multiple operating systems on a single machine.
WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines
o
> do with computers and I learn by reading only.
Believe me, after having worked support for high-speed internet for two
years, the very fact that you know there's stuff you don't know puts you
ahead of the curve. =)
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
&
t had a conniption fit when I realized that the new
card I had with my brand new system didn't want to run under Linux.
I thought that the 3c59x module was backported from 2.4.x to 2.2.19 and
later, but I can't say for sure.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
&q
omehow something got altered
by accident routing wise, but it SHOULD show up in the routing tables
(something like a 10.10.13.0 255.255.0.0 would REALLY confuse the
routing...) in at least some form.
This is an interesting problem (for me, at any rate, probably very
frustrating to you) so I'll bang my
OSPF. I
personally would rather use default route/weighted route methods rather
than OSPF unless there's a pressing need to do so - such as the two
routers mentioned happen to be in totally different locations
topography-wise. Even then, it could be sticky.
Not much help at all, I know, but
thread&tid=128
Note that at present, this is a local root hole, with a possibility for it
to be a remote root exploit - think they're still digging on that.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
"We know what deterrence was with 'mutually assured destructio
e isn't the right one for the kernel you've got
running, actually.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
"We know what deterrence was with 'mutually assured destruction' during
the Cold War. But what is deterrence in information warfare?" -- Br
On Tue, 30 Oct 2001, Simon Bolduc wrote:
> Doing a search on http://standards.ieee.org/regauth/oui/index.shtml for just
> the first 3 hex parts of the MAC indicated that it belongs to HP -
Ah! There's the issue. Yeah, put it in with spaces. Never mind, I'm stupid
today. =)
On Tue, 30 Oct 2001, Michael D. Schleif wrote:
> Yes -- it turns out that mac's beginning with:
>
> 00 30 c1 d8
>
> at least in this case (3 specimens), are HP switches.
Cool, thanks.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
"We know
acturer on that NIC card? I did a search
for the first three at standards.ieee.org and it came up blank, so I'd be
interested in knowing if you've got the info available and easily to hand.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
"We know what deterrence was w
disk could do the same.
>
> We're working on it. :)
Aye, we are. Would be nice if we had a free floppy-bootable device that
worked on hardware totalling about $50 that could do the same job as a
Cisco Pix firewall costing a hundred times as much. =)
--
George Metz
Commercial Routing Engi
it should be just a matter of
adding an ACCEPT rule from 10.0.0.138/32 at the beginning of the IPChains
filter list. That should take care of the issue, yet still leave the
default Martian rules in place.
As for syntax, that I'm not too sure on since I'm weak on IPChains.
--
George Metz
63 matches
Mail list logo