Re: [PLUG] Troubleshooting DHCP...
Galen Seitz wrote: > 10BASE-T and 100BASE-TX only require 2 pairs. There was a version of > 100Mbit that required 4 pairs(100BASE-T4), but I believe it is > obsolete. Anyone still using 100BASE-T4 would almost certainly be > aware of that fact. > http://en.wikipedia.org/wiki/Ethernet_over_twisted_pair > http://en.wikipedia.org/wiki/Fast_Ethernet#100BASE-T4 > > Gigabit ethernet(1000BASE-T) requires 4 pairs.). In its typical configuration, 100BASE-TX uses one pair of twisted wires in each direction, providing 100 Mbit/s of throughput in each direction (full-duplex). Which means that the link could be good, but the cable wouldn't support full-duplex mode. So in theory the NIC might support full-duplex but the Ethernet cable w. a broken wire doesn't. And then you have a duplex mis-match between the 2 end devices. You drop in the 10-base-T hub, which most likely doesn't support auto-neg / full-duplex and it magically works. If I recall, the Orig Poster said he did see Ethernet errors... ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
I believe the following may useful and/or interesting: 100BASE-CX (STP) is only spec'd for 25 meters (82 feet). (Same for 1000BASE-CX, 25 meters.) 100BASE-TX uses two pairs UTP, Cat5 or better, 100 meters. 1000BASE-T needs four pairs UTP, also Cat5 or better, 100 meters. 1000BASE-TX uses two pairs UTP but requires Cat6, 100 meters. Note that most gigabit stuff is 1000BASE-T. I believe the original plan was that two pairs would be cheaper than four but 1000BASE-T won out due to compatibility with legacy wiring, which pretty much is four pairs anyway. NealS ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
Patrick J. Timlick wrote: > I think 10Mbit ethernet requires 2pairs, 100mbit requires 4 pairs, perhaps > you broke a pair or two during installation. > > See http://www.derose.net/steve/guides/wiring/#phone . > > I used this guide to run reliable 10Mbit Ethernet and an analog phone line > through 8 pair Cat 3 buried under a driveway. My run is only about 50 feet, > however. 10BASE-T and 100BASE-TX only require 2 pairs. There was a version of 100Mbit that required 4 pairs(100BASE-T4), but I believe it is obsolete. Anyone still using 100BASE-T4 would almost certainly be aware of that fact. http://en.wikipedia.org/wiki/Ethernet_over_twisted_pair http://en.wikipedia.org/wiki/Fast_Ethernet#100BASE-T4 Gigabit ethernet(1000BASE-T) requires 4 pairs. -- Galen Seitz gal...@seitzassoc.com ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
Patrick J. Timlick wrote: > I think 10Mbit ethernet requires 2pairs, 100mbit requires 4 pairs, perhaps > you broke a pair or two during installation. > > See http://www.derose.net/steve/guides/wiring/#phone . > > I used this guide to run reliable 10Mbit Ethernet and an analog phone line > through 8 pair Cat 3 buried under a driveway. My run is only about 50 feet, > however. > > -- Patrick Timlick > +1 Patrick's theory! ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
I think 10Mbit ethernet requires 2pairs, 100mbit requires 4 pairs, perhaps you broke a pair or two during installation. See http://www.derose.net/steve/guides/wiring/#phone . I used this guide to run reliable 10Mbit Ethernet and an analog phone line through 8 pair Cat 3 buried under a driveway. My run is only about 50 feet, however. -- Patrick Timlick Cat5e On Wed, Dec 30, 2009 at 3:46 PM, Joe Pruett wrote: > > On a hunch after verifying that it still works when I use a 10Mbit hub, > > I replaced the hub with a crossover coupler. That is, I used something > > that is almost equivalent to a 100Mbit hub with 2 ports. It isn't > > working where I'm seeing the same symptoms. > > > > What could be making this work at 10Mbit but not at 100Mbit? The first > > line in the first building is 100' of Cat5e UTP cable, the second line > > is 90' of at least Cat5 STP cable. > > > > I'll have to check the lights on both ends. It just seems odd, I have > > a Netgear FA311 nic on both ends so they should be able to talk to each > > other. > > > > I have a new Netgear switch, but I don't want to open it unless I can't > > get this working. > > > > I wish I had some way of verifying that the line can operate at > > 100 Mbps. > > > > I'm wondering if the Cat5 STP cable is having problems. > > stp is actually more of a problem for long runs than utp is. do you have > the shield grounded at just one end? if not, you could have a lovely > ground loop going on. > ___ > PLUG mailing list > PLUG@lists.pdxlinux.org > http://lists.pdxlinux.org/mailman/listinfo/plug > -- p.j.timl...@ieee.org www.timlick.com 503-476-3119 10990 NE Paren Springs. Dundee OR 97115 ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
> On a hunch after verifying that it still works when I use a 10Mbit hub, > I replaced the hub with a crossover coupler. That is, I used something > that is almost equivalent to a 100Mbit hub with 2 ports. It isn't > working where I'm seeing the same symptoms. > > What could be making this work at 10Mbit but not at 100Mbit? The first > line in the first building is 100' of Cat5e UTP cable, the second line > is 90' of at least Cat5 STP cable. > > I'll have to check the lights on both ends. It just seems odd, I have > a Netgear FA311 nic on both ends so they should be able to talk to each > other. > > I have a new Netgear switch, but I don't want to open it unless I can't > get this working. > > I wish I had some way of verifying that the line can operate at > 100 Mbps. > > I'm wondering if the Cat5 STP cable is having problems. stp is actually more of a problem for long runs than utp is. do you have the shield grounded at just one end? if not, you could have a lovely ground loop going on. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Mon, Dec 28, 2009 at 3:43 PM, Michael Robinson wrote: > > Trouble is, the other end doesn't boot if this doesn't work. I was > seeing packet errors on the far end though before I tried to reformat > the etherboot disk. Can you boot that machine off a "Live" CD? It would give you a lot more tools to work with on that end. Steve D... -- "Every perception is a gamble" Robert Anton Wilson ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Mon, Dec 28, 2009 at 3:43 PM, Michael Robinson wrote: > > On Mon, 2009-12-28 at 15:28 -0800, Mike Connors wrote: > > Michael Robinson wrote: > > > I'm wondering if the Cat5 STP cable is having problems. > > If I were working on this problem, the 1st thing I'd do is use the > > ethtool command > > to look at the physical Ethernet statistics on both sides. If there's > > any problems at the physical layer > > you'll see them reflected in error counters. > > ___ > > PLUG mailing list > > PLUG@lists.pdxlinux.org > > http://lists.pdxlinux.org/mailman/listinfo/plug > > Trouble is, the other end doesn't boot if this doesn't work. I was > seeing packet errors on the far end though before I tried to reformat > the etherboot disk. > > So boot the other end, then switch the hardware. -wes ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Mon, 2009-12-28 at 15:28 -0800, Mike Connors wrote: > Michael Robinson wrote: > > I'm wondering if the Cat5 STP cable is having problems. > If I were working on this problem, the 1st thing I'd do is use the > ethtool command > to look at the physical Ethernet statistics on both sides. If there's > any problems at the physical layer > you'll see them reflected in error counters. > ___ > PLUG mailing list > PLUG@lists.pdxlinux.org > http://lists.pdxlinux.org/mailman/listinfo/plug Trouble is, the other end doesn't boot if this doesn't work. I was seeing packet errors on the far end though before I tried to reformat the etherboot disk. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
Michael Robinson wrote: > I'm wondering if the Cat5 STP cable is having problems. If I were working on this problem, the 1st thing I'd do is use the ethtool command to look at the physical Ethernet statistics on both sides. If there's any problems at the physical layer you'll see them reflected in error counters. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On a hunch after verifying that it still works when I use a 10Mbit hub, I replaced the hub with a crossover coupler. That is, I used something that is almost equivalent to a 100Mbit hub with 2 ports. It isn't working where I'm seeing the same symptoms. What could be making this work at 10Mbit but not at 100Mbit? The first line in the first building is 100' of Cat5e UTP cable, the second line is 90' of at least Cat5 STP cable. I'll have to check the lights on both ends. It just seems odd, I have a Netgear FA311 nic on both ends so they should be able to talk to each other. I have a new Netgear switch, but I don't want to open it unless I can't get this working. I wish I had some way of verifying that the line can operate at 100 Mbps. I'm wondering if the Cat5 STP cable is having problems. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Mon, Dec 28, 2009 at 1:22 AM, Michael Robinson wrote: > > No VLANS are in use. I am running DHCP 3 server on CentOS 5.3 listening > on an actual Netgear FA311 network card with an actual Cat5e UTP cable > running 100' into the attic where at this point I go to a 10BaseT hub > and from there I continue on with a Cat5 STP cable outside, underground, > back above ground, and into another building where I terminate at > another FA311 network card in my diskless server. I tried to replace > the Dlink hub with a Netgear DSS5+ 5 port switch. I guess my switch is > faulty. It seems unlikely that the cables themselves didn't handle > 100BaseTX transmission okay. Obviously there was some networking from > the diskless client to the server at 100BaseTX, but somehow the switch > broke DHCP. Put the slow dumb hub back in, works just fine. Trouble > is, I want the higher speed because I'm extending with 802.11g outdoor > access points. I'm thinking of getting a cable joiner that also acts > as a crossover device and using that instead of a hub or switch. In my google search and a search of Netgear's web site I don't find a DSS5+5. I do find a D-Link DSS5+5. Also the link lights on the switch are for just one set of wires. You will need to check the link status on both ends. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
Michael Robinson wrote:I tried to replace > the Dlink hub with a Netgear DSS5+ 5 port switch. I guess my switch is > faulty. It seems unlikely that the cables themselves didn't handle > 100BaseTX transmission okay. Obviously there was some networking from > the diskless client to the server at 100BaseTX, but somehow the switch > broke DHCP. Put the slow dumb hub back in, works just fine. > It is possible that the switch is faulty, however my experience with auto-negotiation is that each vendor (hub, switch, network card) interpret and implement the N-WAY standards differently. Does the D-Link hub support auto-negotiation? That could explain why the hub works and the switch doesn't. If the hub doesn't support auto-neg then both ends of the connection will fall back to 10mbps half-dup and all is well. If auto-negotiation isn't configured the same on both the switch port and NIC then there will be duplex mismatch errors. A telltale sign of duplex mismatch is Ethernet errors such as Frame Check Sequence errors, late collisions, and runts. You can use the 'ethtool' command to look at Ethernet interface error counters. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Mon, 2009-12-28 at 01:12 -0800, Mike Connors wrote: > Michael Robinson wrote: > > I wonder if being a switch it was blocking the dhcp replies? > Switches don't block broadcast protocols like DHCP. Routers do and if > DHCP traffic > has to traverse a different network than a DHCP proxy must be configured. > > However, there are a couple of cases where dhcp requests/replies could > be inadvertently > blocked. > > 1. The DHCP client and dhcp server are on different VLANs. > > 2. Spanning Tree port configuration is another. With regular spanning > tree (8021.d), > the switch port could be in blocking mode while the spanning tree is > being mapped out and > therefore any DHCP traffic will not pass thru the port. As best practice > we always > configured the switch ports with rapid spanning tree (802.1w), so that > spanning tree convergence > happened faster and didn't impact dhcp requests. No VLANS are in use. I am running DHCP 3 server on CentOS 5.3 listening on an actual Netgear FA311 network card with an actual Cat5e UTP cable running 100' into the attic where at this point I go to a 10BaseT hub and from there I continue on with a Cat5 STP cable outside, underground, back above ground, and into another building where I terminate at another FA311 network card in my diskless server. I tried to replace the Dlink hub with a Netgear DSS5+ 5 port switch. I guess my switch is faulty. It seems unlikely that the cables themselves didn't handle 100BaseTX transmission okay. Obviously there was some networking from the diskless client to the server at 100BaseTX, but somehow the switch broke DHCP. Put the slow dumb hub back in, works just fine. Trouble is, I want the higher speed because I'm extending with 802.11g outdoor access points. I'm thinking of getting a cable joiner that also acts as a crossover device and using that instead of a hub or switch. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
Michael Robinson wrote: > I wonder if being a switch it was blocking the dhcp replies? Switches don't block broadcast protocols like DHCP. Routers do and if DHCP traffic has to traverse a different network than a DHCP proxy must be configured. However, there are a couple of cases where dhcp requests/replies could be inadvertently blocked. 1. The DHCP client and dhcp server are on different VLANs. 2. Spanning Tree port configuration is another. With regular spanning tree (8021.d), the switch port could be in blocking mode while the spanning tree is being mapped out and therefore any DHCP traffic will not pass thru the port. As best practice we always configured the switch ports with rapid spanning tree (802.1w), so that spanning tree convergence happened faster and didn't impact dhcp requests. ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Sun, Dec 27, 2009 at 10:55 PM, Michael Robinson wrote: > On Sun, 2009-12-27 at 21:13 -0800, Daniel Johnson wrote: > > > The other diskless machine on another nic off of the same server, dodo, > > > boots just fine by the way. Is something obviously wrong in my dhcp > > > config file or am I looking at a problem with the switch? DHCP worked > > > when I was using a 10BaseT hub which makes me wonder if the line can't > > > support 100baseTX. > > > > It could be a problem with autoconfiguring the speed, and duplex of > > the connection. I've had two switches decide on different modes, and > > thus get a broken network. When one side is a dumb 10 megabit hub it > > doesn't try to negotiate, so you never have to worry about that > > problem. If you can force it into different modes try that. > > I forced the server to half duplex using ethtool, but I can't force the > client to go to half duplex. Put the 10Mbit dumb hub back in, dhcp > worked like a charm and the machine network booted no problem. I wonder > if being a switch it was blocking the dhcp replies? I should in theory > be able to pull the hub and replace it with a crossover coupler. I'm > starting to think that the switch I was trying to use is bad or maybe > I need a 100 megabit per second hub instead. Thing is, I use another > DSS5+ switch with a linksys wireless WAP11 plugged into it that > configures via dhcp no problem. I'm not sure what the issue is here. > > Just to double check, 190' is within the range limitations for > 100BaseTX, right? > > ___ > PLUG mailing list > PLUG@lists.pdxlinux.org > http://lists.pdxlinux.org/mailman/listinfo/plug > Yes, up to 300' in the accepted range. I don't think Daniel nor was I referring to the client/server but the switch itself. Most switches are auto sensing the link speed and do just fine, without knowing what kind of switch it is and what capabilities it has, it's hard to say what count be causing the problem. Could be VLAN's, could be auto sense not working, could be an ACL .. but without knowing what hardware it is, hard to say. Drew- ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Sun, 2009-12-27 at 21:13 -0800, Daniel Johnson wrote: > > The other diskless machine on another nic off of the same server, dodo, > > boots just fine by the way. Is something obviously wrong in my dhcp > > config file or am I looking at a problem with the switch? DHCP worked > > when I was using a 10BaseT hub which makes me wonder if the line can't > > support 100baseTX. > > It could be a problem with autoconfiguring the speed, and duplex of > the connection. I've had two switches decide on different modes, and > thus get a broken network. When one side is a dumb 10 megabit hub it > doesn't try to negotiate, so you never have to worry about that > problem. If you can force it into different modes try that. I forced the server to half duplex using ethtool, but I can't force the client to go to half duplex. Put the 10Mbit dumb hub back in, dhcp worked like a charm and the machine network booted no problem. I wonder if being a switch it was blocking the dhcp replies? I should in theory be able to pull the hub and replace it with a crossover coupler. I'm starting to think that the switch I was trying to use is bad or maybe I need a 100 megabit per second hub instead. Thing is, I use another DSS5+ switch with a linksys wireless WAP11 plugged into it that configures via dhcp no problem. I'm not sure what the issue is here. Just to double check, 190' is within the range limitations for 100BaseTX, right? ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
> The other diskless machine on another nic off of the same server, dodo, > boots just fine by the way. Is something obviously wrong in my dhcp > config file or am I looking at a problem with the switch? DHCP worked > when I was using a 10BaseT hub which makes me wonder if the line can't > support 100baseTX. It could be a problem with autoconfiguring the speed, and duplex of the connection. I've had two switches decide on different modes, and thus get a broken network. When one side is a dumb 10 megabit hub it doesn't try to negotiate, so you never have to worry about that problem. If you can force it into different modes try that. -- teknotus Take Notice (503) 409-1735 ___ PLUG mailing list PLUG@lists.pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Troubleshooting DHCP...
On Sun, Dec 27, 2009 at 8:54 PM, Michael Robinson
wrote:
> I'm trying to replace a 10BaseT DLink hub with a 100 Base TX DSS 5+
> switch. DHCP doesn't work anymore.
>
> The server is on one end connected via an FA311 netgear card. From
> there, a 100' Cat5e cable runs to the switch in the attic. Then I
> go to Cat5 STP cable for an additional 90' outside, underground,
> and back above ground into the next building. The diskless machine
> that boots via floppy is on the far end. A 190' run made up of a
> 100' and a 90' run seems to be within the range limitations for
> 100BaseTX. Then again, the cable is shallowly buried and could
> be getting weathered.
>
> I see DHCPDISCOVER and DHCPOFFER, but nothing past that.
>
> Dec 27 20:32:36
> goose dhcpd: DHCPDISCOVER from 00:02:e3:02:46:37 via eth2
> Dec 27 20:32:36
> goose dhcpd: DHCPOFFER on 192.168.4.17 to 00:02:e3:02:46:37 via eth2
>
> Most of my dhcpd.conf follows:
>
> failover peer "dhcp"
> {
> primary;
> address 192.168.5.2;
> port 647;
>
> peer address 192.168.5.1;
> peer port 647;
> max-response-delay 60;
> max-unacked-updates 10;
> mclt 600;
> split 128;
> load balance max seconds 3;
>
> peer address 192.168.5.4;
> }
>
>
>
> ddns-update-style none;
> authoritative;
> allow booting;
> next-server 192.168.5.2;
>
>
>
> # %%% wifi-2 %%%
> shared-network WIFI-2
> {
> subnet 192.168.4.0 netmask 255.255.255.240
> {
>option domain-name "w2.robinson-west.pri";
>option domain-name-servers 192.168.4.2;
>option routers 192.168.4.2;
>option broadcast-address 192.168.4.15;
>option ntp-servers 192.168.4.2;
> }
>
> subnet 192.168.4.16 netmask 255.255.255.240
> {
>option domain-name "w2.robinson-west.pri";
>option domain-name-servers 192.168.4.18;
>option routers 192.168.4.18;
>option broadcast-address 192.168.4.31;
>option ntp-servers 192.168.4.18;
> }
>
> # %%% wifi-2 network segment 1 %%%
> group
> {
> use-host-decl-names on;
> next-server 192.168.4.2;
>
> host dodo.w2.robinson-west.pri
> {
> hardware ethernet 00:02:E3:14:1C:94;
> fixed-address 192.168.4.1;
> option root-path "192.168.4.2:/nfsroot/dodo";
> filename="\dodo.nbi";
> }
> }
> # %%% wifi-2 network segment 1 %%%
>
> # %%% wifi-2 network segment 2 %%%
> group
> {
> use-host-decl-names on;
> next-server 192.168.4.18;
>
> host condor.w2.robinson-west.pri
> {
> hardware ethernet 00:02:E3:02:46:37;
> fixed-address 192.168.4.17;
> option root-path "192.168.4.18:/nfsroot/condor";
> filename="\condor.nbi";
> }
> }
> # %%% wifi-2 network segment 2 %%%
> }
> # %%% wifi-2 %%%
>
>
>
> # %%% local lan %%%
> shared-network LAN-1
> {
> subnet 192.168.5.0 netmask 255.255.255.0
> {
>pool
>{
> failover peer "dhcp";
> range 192.168.5.129 192.168.5.254;
> deny dynamic bootp clients;
>}
>
>option routers 192.168.5.6;
>option domain-name-servers 192.168.5.1, 192.168.5.4;
>option domain-name "robinson-west.com";
>option broadcast-address 192.168.5.255;
>option ntp-servers 192.168.5.2;
> }
>
> group
> {
> ...
> # This is populated by host records...
> }
> # %%% local lan %%%
> }
> ...
> # I define the Internet subnet...
>
> The other diskless machine on another nic off of the same server, dodo,
> boots just fine by the way. Is something obviously wrong in my dhcp
> config file or am I looking at a problem with the switch? DHCP worked
> when I was using a 10BaseT hub which makes me wonder if the line can't
> support 100baseTX.
>
> ___
> PLUG mailing list
> PLUG@lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug
>
Have you verified you have a solid link between switch and hub? Have you
verified the switch sees the hub at 10mb? Auto-negotiate can fail and at
times forced or turned off. Plenty of reasons why it might not work beyond
DHCP.
Drew-
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug
[PLUG] Troubleshooting DHCP...
I'm trying to replace a 10BaseT DLink hub with a 100 Base TX DSS 5+
switch. DHCP doesn't work anymore.
The server is on one end connected via an FA311 netgear card. From
there, a 100' Cat5e cable runs to the switch in the attic. Then I
go to Cat5 STP cable for an additional 90' outside, underground,
and back above ground into the next building. The diskless machine
that boots via floppy is on the far end. A 190' run made up of a
100' and a 90' run seems to be within the range limitations for
100BaseTX. Then again, the cable is shallowly buried and could
be getting weathered.
I see DHCPDISCOVER and DHCPOFFER, but nothing past that.
Dec 27 20:32:36
goose dhcpd: DHCPDISCOVER from 00:02:e3:02:46:37 via eth2
Dec 27 20:32:36
goose dhcpd: DHCPOFFER on 192.168.4.17 to 00:02:e3:02:46:37 via eth2
Most of my dhcpd.conf follows:
failover peer "dhcp"
{
primary;
address 192.168.5.2;
port 647;
peer address 192.168.5.1;
peer port 647;
max-response-delay 60;
max-unacked-updates 10;
mclt 600;
split 128;
load balance max seconds 3;
peer address 192.168.5.4;
}
ddns-update-style none;
authoritative;
allow booting;
next-server 192.168.5.2;
# %%% wifi-2 %%%
shared-network WIFI-2
{
subnet 192.168.4.0 netmask 255.255.255.240
{
option domain-name "w2.robinson-west.pri";
option domain-name-servers 192.168.4.2;
option routers 192.168.4.2;
option broadcast-address 192.168.4.15;
option ntp-servers 192.168.4.2;
}
subnet 192.168.4.16 netmask 255.255.255.240
{
option domain-name "w2.robinson-west.pri";
option domain-name-servers 192.168.4.18;
option routers 192.168.4.18;
option broadcast-address 192.168.4.31;
option ntp-servers 192.168.4.18;
}
# %%% wifi-2 network segment 1 %%%
group
{
use-host-decl-names on;
next-server 192.168.4.2;
host dodo.w2.robinson-west.pri
{
hardware ethernet 00:02:E3:14:1C:94;
fixed-address 192.168.4.1;
option root-path "192.168.4.2:/nfsroot/dodo";
filename="\dodo.nbi";
}
}
# %%% wifi-2 network segment 1 %%%
# %%% wifi-2 network segment 2 %%%
group
{
use-host-decl-names on;
next-server 192.168.4.18;
host condor.w2.robinson-west.pri
{
hardware ethernet 00:02:E3:02:46:37;
fixed-address 192.168.4.17;
option root-path "192.168.4.18:/nfsroot/condor";
filename="\condor.nbi";
}
}
# %%% wifi-2 network segment 2 %%%
}
# %%% wifi-2 %%%
# %%% local lan %%%
shared-network LAN-1
{
subnet 192.168.5.0 netmask 255.255.255.0
{
pool
{
failover peer "dhcp";
range 192.168.5.129 192.168.5.254;
deny dynamic bootp clients;
}
option routers 192.168.5.6;
option domain-name-servers 192.168.5.1, 192.168.5.4;
option domain-name "robinson-west.com";
option broadcast-address 192.168.5.255;
option ntp-servers 192.168.5.2;
}
group
{
...
# This is populated by host records...
}
# %%% local lan %%%
}
...
# I define the Internet subnet...
The other diskless machine on another nic off of the same server, dodo,
boots just fine by the way. Is something obviously wrong in my dhcp
config file or am I looking at a problem with the switch? DHCP worked
when I was using a 10BaseT hub which makes me wonder if the line can't
support 100baseTX.
___
PLUG mailing list
PLUG@lists.pdxlinux.org
http://lists.pdxlinux.org/mailman/listinfo/plug
