If you are setting up a router, get a real cheap butt CISCO router. Shouldn't cost you more than 100 bucks. However messing around VPT trunking is a fun experience and you can learn quite a lot from that.

Does anyone know if Linux DHCP servers can do reservations using a MAC address like Windows servers can? For those who are unfamiliar the reservation lets DHCP give a machine the same IP each time as if its static but set at the server level instead of machine level.

Does anyone know if it is possible to set an IP on my HP 2524 (J4813A) without the console cable? The manual suggests it is not but it seems HP used a propriety ping configuration for the console port so the serial cable I ordered will not work. I can get one in a couple of weeks but until then I wanted to actually play with the switch. I was able to use it last night with no configuration just plugging PCs in for a small LAN but I wanted to dig into the options of it.

Not likely, but how is it set up? Try Telnet?

Last time I dealth with HP switches we used SSH but they already had IP addresses. It isn't a huge deal at this point as it works as a basic switch for the moment and I am getting the adapter next week.

I suppose you are setting it up like this? WAN --RJ45--> home router --RJ45--> switch --console(serial)--> PC or whatever?

A couple weeks ago I got an HP console cable (female serial on both ends) from a former coworker and used the USB to Serial cable I had purchased with the switch to try to console into the switch I bought but its not working. The cable is set to COM1 in my device manager but when I try to connect with PuTTY to the switch it doesn't do anything, no errors but no success either. I can't think what I am doing wrong and I checked with the person I got the cable from and he has no ideas. Anyone else have any suggestions?

I have a 500GB external HDD and 2 computers with Windows 7. Is it possible to make the external drive a shared drive so that both of my computers can access the files on it and dump files to it?

Have you tried fiddling with the speed? http://en.wikipedia.org/wiki/Serial_port#Speed. Looked in the manual? It may tell you the specific requirements needed. Have you seen if there is specific software for it? Sometimes there something proprietary being sent that only their program will pick up on.

Yeah, it's really simple and can be done in a couple of ways

1. The cheapest: If one of the two computers is on all the time, plug it into that computer. Right-click the drive letter and select "Properties". Go to the sharing tab, go advanced sharing, allow sharing. Click the "Permissions" button on the same window and allow all permissions. To disable the password requirement you will need to go in Advanced Sharing Settings (Control Panel\All Control Panel Items\Network and Sharing Center\Advanced sharing settings) and turn off password protected sharing.

2. The Simplest: Get a Plug computer like a http://pogoplug.com/devices. Dead-simple to set up, but limited in capabilities and requires proprietary software to connect to it (even on the same LAN) short of hacking it.

3. The most expandable: Build yourself a NAS. You can expand the space as you need it and add other features with ease

A simple ~$150 NAS:

http://www.newegg.com/Product/Product.aspx?Item=N82E16813121442

http://www.newegg.com/Product/Product.aspx?Item=N82E16811147131

http://www.newegg.com/Product/Product.aspx?Item=N82E16820148192

http://www.newegg.com/Product/Product.aspx?Item=N82E16812200294

http://www.newegg.com/Product/Product.aspx?Item=N82E16820139246

+ *nix distro of your choice (like the crafted-just-for-this-purpose http://www.freenas.org/)

You can get a cross-over version of the standard CAT-5 ethernet cable to connect two devices without needing an intervening hub, switch or whatever: they're rather handy. Dunno about USB: "possibly", but I do all my networking with CAT-5.

Yes, they can. I use ISC DHCP (albeit on my FreeBSD server) which does just this.

Edit: http://www.isc.org/products/DHCP/.

Edit 2: example configuration, to be less obtuse about it; it has two fixed MAC address allocations and a general pool:

ddns-update-style none;log-facility local7;subnet 192.168.1.0 netmask 255.255.255.0 { use-host-decl-names on; option subnet-mask 255.255.255.0; option broadcast-address 192.168.1.255; option routers 192.168.1.254; option domain-name "vssnet"; option domain-name-servers 192.168.1.1; next-server 192.168.1.1; group {  filename "/pxelinux.0";  host toaster {   hardware ethernet xx:xx:xx:xx:xx:xx;   fixed-address 192.168.1.7;  }  host aga {   hardware ethernet xx:xx:xx:xx:xx:xx;   fixed-address aga.vssnet;   option root-path "teapot:/ws/aga";   #option root-options "nfsv4,rdirplus,hard,intr";   filename "aga/boot/pxeboot";  } } pool {  default-lease-time 600;  max-lease-time 7200;  range 192.168.1.189 192.168.1.220; }}

Things like filename, root-path and so on you can probably ignore since they're for net-booting Linux systems (and discless Linux is frustratingly lethargic anyway, so I'm probably going to ditch that as an idea.) 192.168.1.1 is both my DHCP and DNS server, just to be clear.

I know Putty works as we used it this past summer where I worked and I have tried both speeds of 9600 and 115200 and accoridng to the manual it will detect the speed and adjust accordingly so it supports a pretty good size speed range so I am confused as to what the issue actually is.

I wouldn't rely too much on automatic speed detection: my experience is that it's not terribly reliable. Ideally, you'd need to find out what's the actual speed setting as well as the byte size, number of stop bits, parity and so on; though I've found the most common problem with serial connections not responding is DTE/DCE incompatibility. And that's something that's always next to impossible to predict, except that it's most likely to be one that doesn't work with the cable you have to hand. If you can find a handy "gender bender" plug that swaps the transmit and receive pins, it might give some enlightenment. (Or it may not: sadly, there are so many parameters for such a simple connection type...)

Edit: actually that's the plug that's a male-female adapter or vice versa, but at least I said what I meant! Blimey, the wrong plug type was also a regular nightmare...

Do you need the pool statement or can you just place it as options and range within the subnet definition? I am also thinking of setting up a Linux domain for personal use plus experience and should I make the name unique or if it is all internal it doesn't matter?

It depends; personally I find things easier to manage if they have fixed addresses, and I think that pools are really only necessary for offices with a large number of itinerant staff. That's really just a hold-over from my first attempts to get my home network together, and my other half's laptop still uses it to obtain its address, but if I was doing it from scratch, I'd just give it its own address. It's not like there aren't enough of them for a typical home network! I think giving each client its own entry provides good flexibility if you have a small number of them, but it'd quickly become an administrative headache if there are lots, in which case defining as many generic options as possible would be a better idea.

In terms of the range being defined outside of the pool statement, I don't know: I'm guessing not since that seems to be its point, but finding out for sure would involve reading the documentation, and my concentration is terrible at the moment...

Regarding names, I'm not quite sure what you're getting at there: I also run my own DNS server, so all my local addresses have unique names (including those defined in the pool) which is generally good practice.

I justs have my subnet declaration followed by the options for routers and broadcasts and then a range statement not as part of a pool and it works fine since I only have one static host set.

Last night I ran into a weird issue which I also had at work though with slightly different parameters. I was building a firewall using iptables and when I set the policy of the INPUT chain to drop even if I have rules specifying to allow certain traffic all gets stopped. I thought the policy was there incase no rules got matched then it drops the packets. Can anyone shed some light on this issue? It also seemed to have happened when I put a deny all at the bottom of the input chain which was also weird since the allow rules were above it.

I recently finished a practice application for subnetting. It is written in Python 3.2 but I am wondering if anyone who is good with that sort of thing might be willing to test it. I plan to have a couple of my former teachers to check it out as well but if anyone else wants to give it a look it would be helpful.

Does anyone know if it is possible to use a Cisco Pix 506e as a DHCP server but not give out a gateway address? I have DHCP set up on my Pix but it hands out the gateway as its address, and when I use it on a wired connection after connecting to a wireless network I cannot use my WAN through the wireless since the wired gateway doesn't exist.

For questions related to specific hardware rather than general networking questions I'd say go to http://serverfault.com/ (http://superuser.com/ if this is for home networking) or http://community.spiceworks.com/channel/show/3-networking (AKA: The Facebook of IT). Both places are pretty friendly and the chances of someone having the same networking gear as you are greatly increased than on here.

So, I have been bored recently and decided to build a router out of a few old system's parts.

http://i438.photobucket.com/albums/qq102/Starforce9/Router2.jpg
3Ghz Pentium 4 (hyperthreaded), 2GB RAM, 160GB 7200RPM IDE drive
4x1Gbps Network cards (3 dedicated, 1 onboard) (PCI based)
1x100Mbps NIC (one dedicated)


Both my WAN and the LAN on Gbps NICs.
I am using PFsense (2.0.1)
When transferring large files from my 192.168.1.0 network to my 172.16.2.0 network I only get around 300-400mbps. CPU usage is about 50-65% during transfer. RAM usage is below 10%.

Would this be a bottleneck from the PCI bus? I dont have Clam AV enabled.

When setting up DNS a domain name is nessesary and I am wondering if I should base my choice for it on availability of the actual name from a site llike GoDaddy if I don't plan to have any publicly available resources? If I choose not to should I still use a .com in the zone name or would a .local suffice as the suffix?

On a related note I am interested in naming convention opinions. At my job we use the company initials then a hyphen and an abbreviation of the system function or department and if there is more than one including a number. I think this is a good idea but for the domains I have in mind it may not work as intended.

I'm not sure offhand if ".local" is officially sanctioned, but it's safer than inventing a domain since Murphy's Law states that one day you'll need to access the real one. Back when TLDs were strictly limited I created my own, but now it looks like they're up for grabs too it may not turn out to have been such a good idea, however obtuse. I've mulled over creating a subdomain of one of my "real" domains for my behind-the-firewall network which would probably be the most sensible choice but I never got round to it for some reason.


Indian curries and variants thereof. You know it makes sense.

If it's not accessible on the Internet, you can use any fake TLD you want that's meaningful for you. It is important to make sure your domain doesn't conflict with any real domain, as your local domain takes priority. This is why .local is often used as the TLD, as it guarantees that your domain isn't going to conflict with any public domain. With the change by ICANN to allow almost any gTLD for the right price, it's pretty important to stay with .local for unregistered domains I think (since ICANN has guaranteed .local is not for sale).

So as long as I don't plan to take any resource public it doesn't matter if the domain is up for sale and I should use the .local? I just wanted to verify otherwise I will need to adjust the domain I chose since the .com is backordered on GoDaddy but it looks like nothing actually exists at the web address. I could use a .us though which should be fine as well.

You can really call it whatever you want. The only problem is if someone else registers a domain of that name and you need to access it, but otherwise, it really doesn't matter (well, occasionally SNMP software likes to make a point, but apart from them...)