Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 10 years ago.
Improve this question
Bit of a random question (and I know a lot of the technology invovled is not described) but are there any networking gurus out there that can answer what is probably a simple M$ networking question that has me baffled?
Connected my laptop to the corporate network today in the same manner I do eveyday. I was allocated an IP address as per normal but unlike normal, I got no connectivity to the internet.
After a reboot and a ipconfig /release /renew the situation was the same. IP address but no connectivity.
So I called in the Techs who replaced my ethernet cable between my machine and the network port and everything sprung into life -- full connectivity restored.
My question: how did it get an IP address using that broken cable?
Most probably, your laptop assigned itself the last IP it was given by DHCP when it failed to get one from the DHCP server.
Which IP did you get? Windows computers get assigned dummy IPs when they do not get one from a DHCP server, starting with 169 if I recall correctly.
There's too many variables to keep a good educated guess. The best guess I can come up with the facts you've given is this:
The cable wasn't broken when it was initially plugged in. Maybe you were able to get an IP address before the cable broke.
As MrValdez already said, there are really to many unknowns to make a good guess. But one thing to keep in mind is, that just because one event follows another it is not necessarily caused by that event. This is also known as "Correlation does not imply causation".
It might well be, that the time to replace the cable was long enough that the real problem (router, ISP or some other thing along the path) could be solved by someone else (without knowing about your problem) or just vanished on its own.
If the cable was broken, you would get an error network "Not Connected" in the icon in the system tray. If you network connection was up, means the cable was OK.
If the cable was not of the right type, then you could have had connection up, but Ethernet negotiated to 10Mpbs/Half Duplex, which can cause a lot of pain to the use who's not aware of it, including symptoms like no connectivity, dropped packets and hence slow connection, etc.
After that, if you got an IP address beginning 169.184.X.X, that would mean that DHCP failed for whatever reason, and your laptop assigned itself a default configured IP address.
If the IP address you got was what you normally get in your LAN environment (the once with which you usually connect to the internet), then there might be some firewall issues, or something else wrong with your corporate network's router, which the Tech guys might not be too willing to share with you :-)
It's impossible to be sure but here are a few:
Network card driver: I've experienced very random things with HP laptops and ethernet cards. Usually a reboot fixes that. Green connectivity light doesn't always mean 'go', sometimes the drivers are responsible for turning the light on.
Loss of connectivity in the cable or the network plug in the compute.
Duplex negotiation problems between the switch and the computer
Maybe trying gigabit ethernet on a 4 wire cable?
If it was me I'd try an ethernet sniffer (like tcpdump or wireshark) if this happens again, see if you're getting anything in. Ping results would also be nice (to IP addresses, not hosts).
The cable and/or connectors were flaky. DHCP is done using very small UDP packets, and if enough of them are sent, eventually they'll get through. The DHCP retry mechanism is very robust, so I wouldn't be surprised if you lost 95% of the packets, and it was still enough to get an IP.
However, once you start sending larger packets for TCP/IP transactions you'll find that the 95% loss drops enough packets that even with durable TCP/IP, it's too tough to get anything useful through.
Related
My internet stops sending outgoing packets a lot of times during the day, and it lasts 8 to 13 seconds.
Example: in Discord I keep listening others, but they can't hear me. Other softwares have the same behavior.
In my house there are 4 routers, and 1 ISP Modem, all in the same subnet.
All routers have DHCP turned off, and the Modem is responsible por the DHCP.
Modem is 192.168.0.1, and the 4 routers from .2 to .5.
I would like to know if this could be a DHCP configuration problem, or a ISP problem! Any ideas?
I tried connecting my PC to all routers, and even the Modem, and the problem happened in all tries.
You have done your troubleshooting well. By observing the problem without any routers in the mix, you've shown that your trouble is in the modem or, more likely, the ISP network behind it.
One more thing to rule out: completely disconnect your routers from your modem, switch them off, and repeat your test directly connected, with Ethernet not WiFi, to the modem.
If you still have the problem you have proven the problem is with your machine, the modem, or the ISP network.
If the problem goes away in that situation it's possible that some other machine somewhere on your network is occasionally blasting data at the modem.
Four routers sounds like an overly complex setup. The typical modem connects to only one router. Routers actually route packets from an upstream network (your ISP) to a downstream network (your building's LAN). If there's confusion about which router is supposed to do that, frankly it's amazing that anything works. If three of your four are configured as WiFi access points that's OK.
The typical modem doesn't offer DHCP service to its downstream LAN network, because it doesn't know about a network behind it, only one device, your router.
Some ISPs provide combination modem / router boxes. Maybe that's what you have. You may want to review your network configuration.
Try tracert command in the command line to see if your packets are travelling through all these routers. Maybe it will solve the problem or at least displays if the problem is caused in the routing. Or try to see if the problem is caused in a wrong default gateway to the device.
I've got an embedded system that just sits their waiting for incoming TCP/IP comms. It's got a static IP address. Do I need to actually set a subnet mask and gateway address as the system NEVER talks out other than when it is talked to?
If I do, please explain why.
Thanks,
Stuart
You can sometimes cheat and do such a thing with IP. The mac addresses should be there, assuming no broadcast, and the ip header has the source/destination, you can flip both around and the ports. You would still need to respond to the ARP looking for you, but the cheat is not sending an ARP looking for them basically ignoring the timeout. To do it correctly you need to keep a table locally and a timer and if you have never seen that ip address or the last time you saw it was more than a minute ago or whatever your timeout is, then you need to ARP for it to be able to respond. TCP is a much larger pain in the ..., with UDP you can do this kind of thing quite easily, swap macs, swap ip, swap ports, fill in the payload, checksum if you want or not, and send it back. Put code in to respond to the ARP. Dont see off hand why TCP would be any different.
Now assuming you are talking about some operating system, embedded or otherwise with a full stack, then that stack is likely going to want that information, but it is operating system/stack specific as to whether you can put bogus numbers in or not for situations where they dont need to be examined to respond to a particular connection. Of course that is not a programming question and this is a programming site, so I assume that is not what you are asking.
I am person trying to learn networking. I understand that this may not be the best first step to take, but I am eager to try and understand how this takes place because it has been tearing up my mind for quite a while now.
My question is, how does a network recognize a device and automatically connect it to the network?(This is assuming, of course, that you have connected to the network previously and are connecting wirelessly.)
Does it store it on the routers side and then look for specific MAC Addresses and then connect it? Or is it stored somewhere on an encrypted file on your computer? Or is it none of these? Please forgive me if I am way off, I am only giving guesses from what I have so-far learned from networking.
'Connect it to the network' doesn't really mean anything, other than just plugging it in, or turning it on in the case of Wifi. What really happens is that the device broadcasts a DHCP request for an IP address, or else It already has an IP address in that subnet. From that point on it is discoverable by ARP, so other hosts in the subnet can send to it.
Closed. This question is off-topic. It is not currently accepting answers.
Want to improve this question? Update the question so it's on-topic for Stack Overflow.
Closed 9 years ago.
Improve this question
All right I admit that it seems like a weird question, and keep in mind (in case you didn't understand it already by the question) I am really not a networking guy, but let me explain:
I work in an office with one (that's right) wall mounted female Ethernet port per desktop PC (about 4 total...all on different sides of the room). This port not only provides Internet access but it also provides access for all desktop PCs to the network for the rest of the building (Alas, I have no direct access to this network, nor I know anything about it).
As I am an external consultant I really don't need/care about the network resources but I do need Internet access for various tasks. I work with a laptop and there is WiFi provided by the company, but as it turns out the desk where I have to sit at is barely in range of this company provided Access Point. Signal is lost often, and I waste hours upon hours trying to get logged back on when I get dropped out. However, my desk is extremely close to one of these wall mounted ports, on which the office's boss computer is plugged into.
So I thought, "Hey, I'll just buy and hoop-up a simple little router to the wall plug and connect both my computer and the boss' computer to the router using good ol' fashioned Ethernet cables (both in a straight-through configuration). Problem solved!"
And it worked like a charm. Or so I thought.
See, the problem is that the boss computer suddenly disappeared from the network (while keeping full Internet functionality, as I did) and as my famously good luck would have it, that computer has some important resources that are needed from the rest of the team/building. Ironically, I was the only one that could see the boss computer on my network because we were connected through the router.
As the local tech support guy explained it, the boss PC could not be found on the network because it had a different IP address which wasn't recognized (these were the actual words he used). Alas, he didn't notice my little router so I unplugged it and fixed the problem for him without him even noticing it.
Maybe the simple diagram below will illustrate better the situation at hand:
Original Configuration:
____ Wall mounted Female RJ-45 Plug
| [] |
'''''' [MyLaptop(with crappy WiFi)]
||
^^^^^^^^^[BossPC]
-----------------------------------------------------------------------------------
Modified Configuration using TP-Link TL-WR740N N Wireless Router:
____ Wall mounted Female RJ-45 Plug
| [] |
'''''' ^^^^^^^^^[MyLaptop]
|| ||
|| ||---------------------------------
|| (Tp-Link TL-WR740N N Wireless Router)
^^^^^^^^^^^^^^^^^^^^ '''''''''''''''''''''''''''''''''''^^^^^^^^^^^[BossPC]
Keep in mind that for now I disabled WiFi completely while configuring the router since I don't need it.
My question is the following:
Is there a way, or a specific set of configuration settings I must use to make it so that I can use this router hard wired as shown above, but that would look "transparent" to the rest of the network? (By "transparent" I mean in a way that my Boss PC can still be viewed by everybody just like it did in the Original Configuration)
There are several ways to make it work as per your description; however, not all of the approaches would work due to the company's security infrastructure. Let's try the simple way first.
1) Disable DHCP on the TP-Link
2) Enable WIFI with your preference settings
3) Plug the wall jack to the TP-Link LAN port (not WAN)
4) Plug Boss's computer to another LAN port
5) Connect your computer to the WIFI with DHCP client enable
Note: In order for you to manage the TP-Link again, you would need to set your computer with static IP address, which must be the same range as the TP-Link IP. Example, if you TP-Link IP is 192.168.1.1 with subnet 255.255.255.0. Then you should set your IP to 192.168.1.10 with subnet 255.255.255.0.
Kindly comment on the post if this does not solve your issue.
Thanks.
There is a lot of history here, so please bear with me.
Our home network used to be fine when we were with Comcast, but we wanted better speed and reliability, so we switched to FIOS. At that point, the Wifi connections from my Raspberry Pis stopped working. I got frustrated after a couple of weeks of trying to discover the cause of the problem and eventually put it aside.
The other day, I started to configure a new Raspberry Pi 4 that I plan to use on my new Sphero RVR. I set it up following the usual directions and the Pi connects to the Internet without problem over WiFi, but I can't ping it from my Windows 10 desktop.
So I started digging. I downloaded the Fing app onto my Pixel 3 Android phone and looked at the network. I immediately noticed that all the devices now have a 10.29.179.xxx addresses. That easily explains why I can't ping from my PC to the Raspberry Pi. They aren't in the same address space.
I did discover that I can ping the Raspberry Pi if I use the IPV6 address. I can even get PuTTY to connect using the IPv6 address, although I was unable to successfully login. I don't know what is causing the login problem, but it's probably something mundane.
I suspect that I can make my Rpi conform to the rest of the network using a static IP address, but that might cause problems if I take the robot somewhere else to demonstrate what it can do. What I would like to understand is why most of my network, using DHCP, is on the 10.x.x.x network and the Raspberry Pis seem to wind up with 192.168.1.xxx addresses.
One further wrinkle. We have two routers, the FIOS router, and a LinkSys WRT1900AC, which is the router on which WiFi is enabled.
I suspect this problem is caused by having the two routers or by something inherent in the way the Raspberry Pi interacts with DHCP.
I would like for my Raspberry Pi to configure with an IPv4 address I can use, no matter what network I am connecting it to, and I'd like to understand why this problem is happening in the first place.
Let me know what additional information you'd like to see.
* Additional Notes *
As to the comment that I have not done research. I spent a week reading through stackoverflow problems and reading up on documentation of DHCP without finding the answers to my question. I do not ask questions idley. I ask questions when I cannot find answers and need the help of people more expert than myself in the areas in which I am having trouble.
I have solved the login problem. It was a stupid password mistake.
So now I can login, over wifi, using the IPv6 address, but not with the IPv4 address.
On the FIOS router, both the 2.4 and 5GHz wifi channels are turned off.
As suggested by Ljm Dullaart, the problem was that there were two DHCP servers on the network. After I turned off the FIOS DHCP, my problems went away.