Networking :: Virtual 'web' For Testing Embedded Devices
Jul 1, 2010
i've put together a slackware box that acts as a "virtual web" to answer pings from devices that ping various pages (such as www.yahoo.com) to test their connectivity. this is for bench testing the devices to verify such complaints as "cannot connect to the internet" or "fails connectivity test". so far, so good, except for a few devices that seem to be also try to load a particular page wirh an embedded browser. this i need help with. currently, all that's needed to return a ping is to look at the DNS requests the device makes and add the requested web address (www.yahoo.com for instance) to the hosts file. what i need is a redirect to a particular html file based on the address requested (i.e. yahoopage.html for www.yahoo.com, lgepage.html for www.lge.com, etc) any deas on how to accomplish this?
I am in the process of writing a socket program that will need to talk to different computers across the world. I want to set up a virtualized WAN on my computer so i can install the software on a virtual machine, and test it's connectivity in a "real"ish WAN environment.
[URL] Looks promising for setting up the WAN, but i was curious if anybody knew a better way to do it than setting up a bunch of virtual servers, and then another virtual server running WANEM to act as the WAN. I have never been very successful with Virtual Nic's
Now doing some research about performance of Linux switch. I'll start with basic structure of my system: PC1 is connected to PC2 - both have 1 NIC so not much to add.PC2 eth0 is divided in VLAN`s as far as maximum of 4094. Every VLAN has it's own switch (bridge). On PC1 when I run a test I just make VLAN/configure IP address and stream data to IP that's configured for that bridge on PC2.But now I want to add one more interface to each bridge, so I could stream from PC1 through bridge to Virtual interface (just an endpoint out of bridge). I could possibly add one more computer, physical interface to PC1 and make it work, but is there any way to make this work with virtual interfaces, like tun or something? Making just a simple tun interface and adding it to bridge, as you understand, doesn't do the trick, I need some kind of other endpoint that /virtually/ connects to tun interface. Maybe someone has made some kind of script or something? Using Iperf for my tests.
I am looking for tools for static/dynamic code analysis for embedded Linux system development (both device driver and user space apps) with ARM-based processor. We use Eclipse IDE and C++ lanuage for development. Does anybody have recommendation for tools to analyze code complexity? The tools is better to support McCabe complexity metric, however, we may also consider others. Does anybody have recommendation for unit testing?
I'm using fedora 11. I'd like to do some tests with direct IO (raw) devices. I have now raw(1) command. Yum fails to find a package providing raw. There's a /dev/rawctl and raw appears in /proc/devices. Where is it? (rpm?)
We have designed a board with Cirrus Logic(arm) processor, A Flash memory and some other peripherals have been connected to that. While building kernel we have selected MMU support. We have written few custom drivers for keypad,LED,LCD. But I would like to know how virtual memory mechanism can be helpful here even though there is no any hard disk has been connected. Where will be the virtual memory reside.
I'm using the debian testing netinst image (from below url) and trying to install on virtualbox (part of a debian testing vagrant base box build pipeline I've got). This hasn't worked properly for last couple of weeks. Host system is also debian testing.URL....
if I accept all the defaults through the installer (apart from small details like timezone), I end up with a system in which the network does not work. It seems that /etc/network/interfaces has eth0 but the actual network card is called enp0s3, so no network comes up at first boot (despite that the network works fine during the actual install). I can easily fix it after the first boot by editing /etc/network/interfaces and replacing eth0 by enp0s3 and then doing an ifup. This is fine for a workaround but begs the question of why does it fail in the first place.
I have googled a bit and found this thread: [URL] .... which describes a similar issue, and there are other threads out there describing the change.
The udev change seems reasonable enough, but also some package involved in basic debian netinst installation has a bug given that the default install path results in a broken system. Question is, is there a bug already reported (I couldn't find it), or else, which package should I report the issue against? End result should be basic install does not require editing after boot to fix the network.
I am running federa 12 and just recently installed virtual Machine Manager. I've created Virtual machine and installed Windows XP on it. Everything is running OK, as far as I can see. I've installed software to manage and download pictures from Cannon camera, but I am not able to make USB port available to this Virtual machine. I tried to add hardware in VM, but not sure if I should go for STORAGE option, or PHYSICAL HOST DEVICE. I tried second option, and selected USB device. Now I've got UB hub available, but still cannot see anything attached to it (camera in my case, or USB stick with pictures).
I'm noodling around with Ubuntu 10.4.1, latest updates and kernel (2.6.32.24?).Anyway, I run ClamAv as root and it goes fine through almost all of my system (huge amount of it), including several virtual devices, where it hangs on pan0, which has some association with my network (eth0 would be for wired connection, and wlan0 for wireless, and pan0 is listed also, but I'm not at that machine right now, so I can't tell why it shows up. wlan0 is what I use to connect to the internet).Is there an issue for clamAV with virtual devices? Any workaround? I had to terminate the scan after it stayed hung for over 5 minutes on pan0.
i work in consumer electronics, and due to company constraints on network usage, i came up with a "jail box" to test blu-ray players, TV's etc that can connect to the internet. the majority of the complaints the server addresses are mailnly "blu-ray player can't connect to the internet", etc... right now it's set up only to return pings that the devices do to test connectivity (i.e. entries in the etc/hosts file for www.samsung.com, www.lge.com, etc.).
i want to have the box actually provide DNS, samba shares, possibly a "pseudo-you-tube" link to test other functions, but the first one i want to tackle is to have the box providing actual DNS services (even though all DNS queries will after all point to the same ip address) the samba shares i know isn't a big headache, i already have it providing virtual web pages (but with no content as yet), the "virtual you-tube" page i'll tackle later. so how would i go about setting up BIND to do this? keep in mind this is a standalone box, not connected to anything but the device under test.
I am working on a NFS server embedded on a PowerPC plateform (4650EX, 512Ram, 1 GB Ethernet)but i can't mount my exports folders from my client. Here are messages :
-- server side : $ tail /var/log/message Dec 11 23:35:45 canyonlands mountd[1345]: authenticated mount request from 147.138.27.17:620 for /mnt/ssd
I am running Virtual Box Ubuntu Linux 2.6. Connected to my PC is an embedded board running debian 2.6.32. I have ip forwarding on at my virtual box. From my embedded board I can successfully ping www.google.com. I want to download and install some packages but when I try apt-get I get the following: I've flailed away at this for most of the day
When I create virtual ethernet interface and do a ping -I <v_int> <host> the outgoing address is the one of the physical interface and NOT the virtual interface.Is there no chance that trafic will go out with virtual interface address??Incoming trafic is done well i.e. responds to the virtual interface have the virtual address.
My problem is that I have 2 modems and want to check both default gw behind the modems. If I do a "normal" ping both are reachable over default route even the modem which is not the default route will not work because ping goes over the working modem.So I have 2 routing tables and want to route the virtual interface to one modem the other to the other modem
The update that broke things contained some 400 packages.The problem is NOT hardware related. There are 2 other linux installs on the laptop that work the way they should.This is all related to using a wired network connection.The update removed the package network-config and did not replace it. Not sure that is the problem because that package is related to the gui and I can not get an ip in cli.
I am stuck at cli because the new kernel broke nvidia (pretty normal). There was also a new xserver, which tends to break nvidia, and as I recall a new kde. Anyway I am unable to rebuild the nvidia driver because I can not establish a network connection to my network to get the packages needed to fix it.
I would like to test myself if port 5060 is open for UDP packets.
I got a few shell accounts outside of my ISP that run Linux or different BSDs. Does someone know a common program where I could send UDP packets to port 5060 of the IP address that Virginmedia assigns to me and then I would need a program on a PC within my network who reads them _if_ they get through.
Background of all this is that my VOIP phones from two different VOIP providers are offline since last Monday. One service provider confirmed that another VOIP client is online with the same service provider Virginmedia but in a different area.
So, while I don't know how long does it take that someone from Virginmedia can tell me if their port 5060 is open for UDP, I would like to find out myself in the meantime.
My program is working. However, I would like to add in some features. Let see my code first
Code:
if (returnresult<=0) { printf("Connecting.... ");
[code]....
When the ethernet network connected, it will go into the else loop at line no. 13. The system will send status and display the status at control panel screen. If at this point, the ethernet disconnected, it will go back to the if loop at line 1. The system will keep trying to connect to the network and keep loop the if loop until it is connected.
Question now is I need the system to test the connection too after going into the operation loop that is start from the line no. 47. If the system went into the for (int c=0;c<cCOL;c++), it will need to do all the operations as indicated by cCOL times before it exit that loop and go into the previous loop.
I need the system to detect connection everytime an operation is going to happen or ended and if there is no connection, it will trying to connect. Before it is connected, it will not continue the operations. I tried to put the function "Connect (szPort, szAddress)" into each loop. Yes, the operation stop everytime the network disconnected. However, it shows on control panel screen "New Connection...." everytime it execute the statement. In fact, the network never disconnected.
I tried to look at the Connect function code, it is as below, I cannot figure out any statement that can stop the appearance of that "New Connection...". It may be hidden at the library source code, it is too hard for a beginner to modify the library code.
Code:
int Connect (char *szPort, char *szAddress) { /* Set the remote port */ int port; /* port number */
I am trying to do some bandwidth management testing, so I have developed a small script for doing so. When I try to run it it gives the error <error> RTNETLINK answers: invalid argument</error>. I have searched the net, and so far they say it has to do something with the kernel support.
the script: <code> tc qdisc add dev eth1 root handle 1: htb tc class add dev eth1 parent 1: classid 1:1 htb rate 2mbit ceil 2mbit tc class add dev eth1 parent 1:1 classid 1:10 htb rate 600kbit ceil 2mbit prio 2 tc class add dev eth1 parent 1:1 classid 1:20 htb rate 1400kbit ceil 2mbit tc class add dev eth1 parent 1:20 classid 20:100 htb rate 700kbit ceil 2mbit prio 0 #when running this command it gives the error tc class add dev eth1 parent 1:20 classid 20:200 htb rate 700kbit ceil 2mbit pro 1 </code>
I am running this script on a virtual machine that has ubuntu 9.04 installed.
I am trying to set up to serve Linux kernels and filesystems disklessly, and find the startup and test procedure a bit tedious.Can anyone point to a tool or tools to allow me to see what is going on with my DHCP/TFTP/PXE setup, which I'm trying to get working? Ideally, I'd like a tool that can do something like grab an IP from a DHCP/BOOTP server, tell me all about all that it found, and then if it was told to download a file, it could do that and tell me about how that went, and so on. Between the information reported by the tool and the droppings left in logfiles on the server side, I would hope to simplify the configuration process, as well as let me experiment with various options, etc. Even a PXE client tool that I can use to exercise my tftp PXE server configuration from a shell commandline would be extremely useful.If no such tool already exists, would anyone else find something like this useful?
I know pretty nothing about networking'n'stuff. I got this linksys wireless router that was somewhere in a box for a few years ... now I want to see whether I can bring it back to life. I know, that I put some different software on it and that it doesn't have it's original ip assigned any more. But I can't remember which ip that would be.
So for starters I just wanted to put the original software back on to see if that thing still works. But how to do that? First I got no clue how to determine the address it is listening to. Once that'd be figured out I guess I'd have to ssh to it and install this ____.bin file I downloaded somehow.
What I am trying to do is the following: I have many IBM MT-M 6223 towers (Intel Xeon processor) that need to run a certain Fedora 9 configuration. It is working on one box, now it needs to be replicated on all of the others. What is being done is the working box is cloned, and then that image is loaded on all of the others.These boxes have 5 total Ethernet devices. There is one on-board port, and there are two 2-port expansion cards for a total of 5.
There was one initial problem out of the gate that has since been solved (or perhaps not completely). What was happening is that the boxes that received the cloned image had their Ethernet devices misnumbered. Instead of eth0 through eth4, they were numbered eth5 through eth9.This was eventually traced to the following file: /etc/udev/rules.d/70-persistent-net-config. That file contained the original settings from the box that was first cloned. It had eth0 through eth4 at the top with MAC addresses belonging the box that was the source of the cloned image. It looks like this mismatch was detected, and five more entries, numbered eth5 through eth9, got appended to the bottom with MAC addresses belonging to the new box.
Once /etc/udev/rules.d/70-persistent-net-config was deleted, and the machine rebooted, the 5 ethernet devices came up with the correct numbering (0 thru 4). /etc/udev/rules.d/70-persistent-net-config was rebuilt with the correct devname-MAC pairing.
Now for this new problem.... Part of the configuration across all these machines is that only eth3 should have an IP address. It is a static IP address set through the /etc/sysconfig/network-scripts/ifcfg-eth3 file. All of the other devices must NOT have any IP address. They are either not used or used only for raw Ethernet.After clearing out /etc/udev/rules.d/70-persistent-net-config, /etc/sysconfig/network-scripts/ifcfg-eth3 was edited to give the new box its own unique static IP address before rebooting.Sample ifcfg-eth3 file:
I'm looking for some help on how to troubleshoot the network on Ubuntu 10.04 64 Desktop. It's on a Acer Aspire 721. There's a wireless card but the widget in the top right says there is no network device.
Just built a box and once 10.04.2 64 bit was all installed and loaded up I realized instead of showing my Ethernet as connected it had the ol x. So when I clicked on it I got "No network devices available." The mainboard is an Asrock 890GX Pro3, link light completely out.
How can I find all the devices(printers,etc...) and computers along with their OS if any installed that are connected in my intranet (devices/computers may be down here)?
I did this earlier using netstat or nmap not sure what I used and how I did.
Just wondering if there is a way to view all connected devices on my network?? Something similar to view network map in vista?? I know the places network but that only shows me my shares and I want to be able to see everything that is on there ie printers, routers, VOIP, other non-shared computers, etc...
In my home network, My Ubuntu 10.04 LTS (fully up to date) is having trouble being able to ping other computers by their computer names instead of by their IP#s. I've checked over the settings of the router (D-Link that provides wired and wireless connectivity, DNS, Gateway, etc.) and I'm pretty sure it is set correctly. My Ubuntu system is a dual-boot (Ubuntu and XP), the other computers are all XP except for one which is running Windows 2000 Pro. All the other computers are able to ping each other by name, and can ping this one by name when it's in XP or when it's in Ubuntu.
But this computer (when in Ubuntu) cannot seem to ping the other computers by name. Some it can be pinged by name from the Ubuntu system if I include the .local to the name, but others do not. I have checked and compared the network settings on each of the computers to make sure they were set identically (only the computer name is different). All of the XP systems are DHCP configured, The Win2000 is set static and is also reserved at the router. NONE of them have any of the computers listed in a hosts file. I want to avoid editing hosts files if I can.
There are seven computers in all. Only the Ubuntu has difficulty pinging other computers by name. What should I check? What can I set? I would like to know what it is that I'm missing.
I'm trying to set my network interfaces so that they don't get random every boot.(eg assign eth0 to a network interface with a given MAC addr, and eth1 to the other one)I trew in a udev rule (in fact just modified the rules that was automatically generated and set the ethX in it) but the system ignores my udev rule.What am I missing ?Here is all the info :
Code: $cat /etc/udev/rules.d/70-persistent-net.rules # This file maintains persistent names for network interfaces.
I did a forum search and too much came up that was not the answer I was looking for. I have only Linux devices at work and I was wondering how do Linux Administrators connect to Cisco devices via serial console cable. Cisco recommends 'Hyper Terminal' which was a Windows add-on applications licensed to XP. I think Vista and 7 no longer have the license to use Hyper Terminal. Regardless I don't have any versions of Windows and would like to use a serial cable from Com port 1 to the console port on the device. I can't use SSH since some devices don't support SSH and or I have to do a password recovery.
I'm doing a few tests with fedora 15. I'm surprised because they changed the naming way on network devices. eth0 is now called em1.At every fedora I have found the configuration files in /etc/sysconfig/network-scripts, but today I can't find them.
