Ubuntu Networking :: Strange Behavior With Pings From Desktop?
Mar 9, 2011
I've been trying to use MySQL in Ubuntu but I've been having some connection issues and in trying to troubleshoot that, I observed this strange behavior with pings in Ubuntu Desktops inside our network.
- All Ubuntu Servers we have (10.10 and 8.04) behave as expected
- All Windows machines behave as expected
- All 5 Ubuntu Desktops we have, (10.10 and 9.10) exhibit the following unexpected behavior:
If they have a local ip addresses and are connected within the internal network, either to a switch or directly to a router, pings take much longer to be sent. It's not a higher lag, it's the time between each ping that increases. I have to wait for around 5 seconds for each ping, but the time in the ping reply is less than 100ms, also they're all in order so I assume it's the system delaying the sending of each ping.
If I plug a USB 3G modem directly then it works as expected again (~1 sec between each ping sent). Is this an intended feature, that depends on the network environment, or is this something I should look into more closely?
I have an old PC running an older version of linux with two network cards serving as my firewall router. The network cards are netgear 310tx cards. Noticed traffic was not getting routed at times. Figured my switch was going bad, so I hooked up an old hub to keep keys systems communicating while I get a new switch. No problems with hub, but only has 5 ports.
Get new switch (different make/model from old one), install, a couple of my systems have connectivity problems, but others do not. Thinking it was a bad switch, returned it, got a new one just like my old one (which I thought was going bad, but worked for several years). Hook it up, paying attention to my firewall console and notice martian header messages. Network traffic not getting routed at all. The interface for connected to LAN was slow to get started (my firewall can communicate to the Internet, but LAN communication was hosed).
System would "hang" spitting out martian header messages until I disconnected LAN cable. Thinking network card may be bad, replaced it with another version of the same type. Same problems: martian header messages and traffic not getting routed.
HOWEVER, if I remove the switch and use my old hub, no problems. Networking appears to work for all systems connected to hub. Since I and my wife work from home, I'm using the hub right now (with non-essential systems not connected due to limited ports). I can only do any real experimentation at night.
I have no idea why things would work differently between using a hub vs a switch. Could something have gone awry with the linux kernel (2.4.20-46.7)?
I have a small ARM machine running an Arch linux variant, PlugApps. It was set up and working fine at my home, with a standard networking setup -- cable modem and router. I could do pretty much anything I needed to access the net fine.This machine is a remote monitoring system for another location. I shipped it to that location and am now trying to get it set up using another set of hands on site. There it is on the network behind a dsl modem and a router. I forwarded ports from the dsl modem to the router and from the router to the ARM computer.
In both cases the ARM machine is/was configured to use dhcp.So here's the deal:I can ssh into the ARM machine just fine.I can ping hosts on the wan fine (eg: stanford.edu).BUT:wget fails for not being able to resolve a valid address (eg: http://checkip.dyndns.org)the PlugApps package management system (pacman) cannot resolve the repository address the openntp time sync isn't working On the new network, I am able to connect to another client windows PC which has the same associated network info (Bcast address, Mask, etc. - comparing windows "ipconfig" vs linux "ifconfig" outputs) but which can resolve addresses for web browsing, etc.What am I missing? If it can resolve addresses for pings, why can't it work for other address requests? If I can ssh interact with it, why can't it connect to the net for other purposes?
I have recently upgraded to 11.4 and also had to renew an AP (HP ProCurve 10ag). This was several years old but replaced an identical model which had been working well until recently. I thought it would be appropriate to flash newer firmware as the device four releases behind the times. The access point serves partner's XP laptop and my openSUSE 11.4 laptop and after flashing the firmware both machines would make a connection and then break and remake, to the extent that the openSUSE machine became unusable. I thought it was an encryption problem and spent hours changing the AP and Client setups to no avail
After reading a thread here I checked out dmesg and this is what I found:- [78.639247] wlan0: authenticate with 00:1d:b3:4b:ae:14 (try 1) [78.640804] wlan0: authenticated [78.642507] wlan0: associate with 00:1d:b3:4b:ae:14 (try 1) [78.651195] wlan0: RX AssocResp from 00:1d:b3:4b:ae:14 (capab=0x411 status=0 aid=1) [78.651200] wlan0: associated [78.652222] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready [78.652283] cfg80211: Calling CRDA for country: DE [78.674378] cfg80211: Regulatory domain changed to country: DE ..... This went on and on.
I have just installed a testing net install of squeeze on my laptop and I'm having a few problems with my wired connection.
When i try to ping google.co.uk there is a large delay between each result even through the actual ping takes ~30ms. It takes about 5 seconds or more between each ping.
If i use ping -n it works fine. Does anyone know why this is happening?
I've noticed something strange in the behavior of how Ubuntu server obtains it's IP address versus Ubuntu desktop and other versions of linux and Windows. When Ubuntu desktop obtains an IP address from my router that address is retained from one bootup to the next, same behavior as Windows, SuSE, and pretty much any other OS. Ubuntu server on the other hand grabs a different IP address everytime it boots. At first I guessed it was a difference in /etc/dhcp3/dhclient.conf so I replaced it with the dhclient.conf from Ubuntu desktop but no dice, it still grabs a new address every time.
In /etc/network/interfaces Ubuntu server defines the loopback interface and primary network interface, just as it should but Ubuntu desktop does not define the network interface, only the loopback interface. I'm guessing something else controls does this. Desktop appears to be storing previous IP lease information in /var/lib/dhcp3/dhclient.eth0.leases but server doesn't, /var/lib/dhcp3/dhclient.leases is always empty and there's no other files in /var/lib/dhcp3/.
Apparently something isn't writing old lease info there. Without setting a static IP address, how can Ubuntu server be configured to retain the same IP address between boots like Ubuntu desktop?I'm using Ubuntu server 8.04LTS, Ubuntu desktop 9.10 & 9.04, SuSE 11.0, Linux Mint, Windows XP, pfSense 1.2.2 as router.
I am having (seemingly) random trouble with my wired network ever since I installed Lucid. I have no problem getting an ip address from dhcp. However, randomly the computer will boot and although I have an ip address I do not receive any responses for pings on the network nor can I browse the web. If I sudo /etc/init.d/networking restart a few times (or reboot) it will start working. However, restarting the networking services (as mentioned above) again will cause me to no longer receive responses for pings or browse the web.
Furthermore, I have never been able to successfully ping if I manually set an ip address. I have un-installed network manager and I am using /etc/network/interfaces to configure the network. Using Lucid Lynx 64bit on a Dell Precision. I have pasted below the output of a few working commands. When I switch between static ip and dhcp I am commenting/uncommenting the lines shown in /etc/network/interfaces.
I have an asus router running tomato firmware. Through the webGUI I have set up an openVPN server as per the tomato guides. I am using the network manager plug-in in ubuntu 10.04 to connect as a client to the tomato openVPN. All of my certificates seem to be set up and working correctly.
I am able to successfully connect to the openVPN server using the ubuntu client but after that I get nothing. On the client side I am unable to ping the router, any ip addresses inside my private network, the virtual network ip, and web browsing (client) no longer works.
Also, when connected I am unable to ping the client from the private network.
Before connecting
Code: chris@ChrisLaptop:~$ ifconfig eth0 Link encap:Ethernet HWaddr 00:1e:68:90:5e:e9 UP BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0
I'm running KDE on Fedora 13 and was trying to alter some display settings so that the windows wouldn't maximize when dragged to the edge of the screen. I did that, but at one point "Kwin" crashed, and now there's some odd things happening.
1. I get the "drag me" palm when I'm mousing over the desktop and if I click, I can drag the background around as if it were a window.
2. My "maximize" "minimize" and "close" buttons don't work.
3. It appears nothing in the titlebar of windows works: can't drag them, can't shade or do anything "clickable" -- close, min, max, nothing works.
I tried returning the settings to their original state but it didn't help.I tried restarting the computer too, and that didn't help. I'm guessing something crashed and a setting got hard set that shouldn't be. I'm also guessing that if I rename a .kde/something folder it will reset to defaults. What .kde folder should I play with?
Brand new network, just installed, all ubuntu 9.10 boxes connected to a Dell 2206 dumb gigabit switch, and from there to Dlink router to cable modem. Pretty simple. ISP confirms great link out and inbound,
However, http requests can sit for 20-30 seconds (sometimes longer), before content returns, same for pings.
Some thought it might be MTU-related. What is best MTU setting 1500/1492/ ?
I noticed yesterday, after what must have been the most recent updates, that my display now has a dark striped bar running down the right hand side of the display. It runs between the indicator applet and the deleted items bin on my standard Gnome desktop. I know it's not my monitor since I can see both of those items but maximised applications will not fill this area, even though the cursor is visible there.I've tried changing a number of display options but to no avail.
I have a Slackware 13.0 server that is not replying to icmp pings. However, I have double checked that /proc/sys/net/ipv4/icmp_echo_ignore_all is set to 0. The weird thing is I can run nc -v slackware 22 from any computer on my lan and the banner comes up just fine. The nic is on the same netmask and broadcast as all the other computers on my lan. I can log into it as well, it just doesn't respond to pings. I can even ping other computers and get replies just fine when pinging from inside the Slackware server.
I have even added an iptables INPUT rule for -p icmp -j ACCEPT. Although I've never had to do this. I ran tcpdump -vv icmp and I can see ping requests coming into the computer, however slackware simply doesn't respond. What could be causing this? A corrupt arp cache? It's been happening for more than a few hours so I wouldn't think so. How can I view the arp cache table? How do I clear it?
For some reason, the desktop for Fedora 12 has the following closing tag:
Code: </body> <
I have just installed Fedora 12, and I am using GNOME. Even the live disc had the tags. I had assumed that after installing the software to the hard disk I would be rid of the problem. Apparently I was wrong. Google does not seem to bring up a solution. Has anyone else encountered this?
I use Debian Lenny on desktop. On bootup I get this strange message
Code:
Question one: what is hub 2-0:1.0? A utility? A library? A module?
Queston two: what does "enumerate" mean in this context?
Question three: I can't figure out what "port" is meant: a networking port something on the motherboard one of several I/O port/connector/sockets on the back of the computer? (Hope you know what I mean--- you stick USB connectors into them to connect your computer to a USB device, such as a USB mouse.) the one and only networking port/connector/socket on the back of the computer.
Possibly relevant: I sometimes but not always have a USB cable connecting my printer to one of the USB ports/connectors on the back of the computer at boot time. But the message always refers to "port 1", never 2,3,4. due to lack of options for getting on the Net, I use an Actiontek router which demands to be given 192.168.0.1 which would otherwise be taken by my computer, which instead gets 192.168.0.2. (Grasping at straws...)
The computer and system in question is the laptop from my description and openSUSE 11.3 with a different kernel from this repository :[URL]..free to move this thread if it should reside in the beta/prerelease forums section. The IMHO relevant information about my hardware and system is shown below.
The kernel I'm using :
Code: grzes@opensuse:~> Linux opensuse.local 2.6.37.2-1-desktop #1 SMP PREEMPT 2011-02-25 19:59:41 +0100 i686 i686 i386 GNU/Linux The hardware and drivers I'm using : Code: 02:00.0 Network controller [0280]: Atheros Communications Inc. AR9285 Wireless Network Adapter (PCI-Express) [168c:002b] (rev 01)
I've got a machine running Ubuntu Server that is on several VLANs. Each VLAN has its own subnet and the server has an address on each subnet. The switches are set to allow tagged traffic to the server for each VLAN that it is on. Switch ports ending with workstations are given untagged ports on whatever VLAN is appropriate. Workstations are given addresses on a subnet for each VLAN via DHCP. All this works great and hosts on any subnet/VLAN can access the server as normal via its address on that subnet/VLAN.
Accessing the machine by its address on a non-local subnet is where I run into a problem. Inter-subnet traffic has to go through a router, which has been set up appropriately. Running tcpdump on the server and pinging it from a workstation on a subnet, using its address on a different subnet, shows the server receives the ping, but sends no response:
Code: sudo tcpdump -i vlan4 -n tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
At my grandparents house, there is a Wi-Fi network that we can (sometimes) connect to. However, it lately hasn't been allowing our Win 7 laptops to connect. It does let my Wii and my iPod Touch connect with no problems. I've tried restarting, but that doesn't work. We don't have access to the router, so we can't reset that. I would really like this fixed since mobile Safari is not the best browser in the world and WiiNet is even worse..
Note: Neither of the laptops are mine (one is my sister's and one is my parents'). I would never allow a computer of mine to use such an operating system.
I just installed Ubuntu 10.10 yesterday, this is my first experience of Linux so I am a noob and have no idea how to go about troubleshooting this myself. I have been googling and searching forums for ages and have read a lot of threads of people having problems getting connected but have still to find anyone with the same issue I am having.
I managed to connect without any problem and I am not being periodically disconnected (which seems to be a frequent problem) but when I enter an address in firefox or click a link there is a 20-40 second delay before it will take me to the page, but once I am on the page if I click another link it will take me there straight away (as long as I am in the same tab, if I switch to a different window I have the 30 second delay again). Same with downloading files, when I click a download link it takes about half a minute to register that I have clicked, but once it establishes the connection my speeds are the same as they always are. speedtest.net is showing 8mbps, and the last file I downloaded came through at 900kbps once the connection established. It is very irritating since other than that I am loving Ubuntu and the more I read about Linux the more I want to drop my windows 7 altogether (running dual boot just now).
Right then... all DHCP PEGASUS 192.168.0.101 When being pinged from another PC aok, but also pings back on 192.168.0.100? When being pinged from itself using ping 192.168.0.101 (or 100) all is well - when pinged as ping PEGASUS is says following:
64 bytes from localhost.localdomain (127.0.0.1): icmp_req=1 ttl=64 time=0.033 ms 64 bytes from localhost.localdomain (127.0.0.1): icmp_req=2 ttl=64 time=0.044 ms 64 bytes from localhost.localdomain (127.0.0.1): icmp_req=3 ttl=64 time=0.042 ms 64 bytes from localhost.localdomain (127.0.0.1): icmp_req=4 ttl=64 time=0.040 ms ^C64 bytes from localhost.localdomain (127.0.0.1): icmp_req=5 ttl=64 time=0.042 ms
The IP says its 127.0.0.1 (home) WTF is going on>. Why has PEGASUS been assigned 2 ips, and when pinging using HOSTNAME why does it point to home?
I've a question about a connection to a shared folder on a pc behind an ubuntu 10.04 server. The Ubuntu machine itself is behind a Cisco router. Cisco router has a nat 1-to-1 to Ubuntu machine with the static public ip x.y.x.z The Ubuntu machine has one only interfaces eth0 bridged with a virtual tap0 in the virtual br0 interface. This because on the Ubuntu machine is running OpenVPN in bridged mode. I want to reach a private share on the host 192.168.100.2 using the x.y.w.z static IP. Is that possibile? Ubuntu machine has no firewall running (I was not able to run shorewall firewall for now).
I'm getting these strange answers from opera when I ask it what speed for download does it have. I am downloading fedora using ubuntu right now. Why opera tells me it's downloading with almost 200kb/s while the system monitor from ubuntu can't pass more than 110kb/s ...
I need to check the output of a name resolution, but I configured the firewall of the destination to avoid answering to pings. Is there any other way to get a name's IP without using ping? (I mean something like $resolv name.com)
I usually start up my Internet service by typing ./rc.inet1. While its connecting I usually see my routers ip address then I connect and I start browsing and everything is fine. A few times I noticed that instead of seeing the routers address it showed some other address and I could not get online. I try ./rc.inet1 stop and renew but still cannot connect, I reboot and I then see the router and I can connect. I am using slackware13, and opera10 browser.
I opened a thread on the general networking section of the forums to discuss about a problem I am experiencing with a newly installed slackware 13.1 (not 64b) on my dell laptop. Basically, the machine is not getting an IP from the dhcp server on my pfsense router. All other machines on my network are still getting IP's fine, and the pfsense box never changed. I can get an IP from the pfsense box by manually loging as root and using /etc/rc.d/rc.inet1 eth0 restart (note the argument eth0 necessary to get the ip).
User markush at the link below suggested I open this thread here, I think its a great idea since I believe it is a slackware problem. he also recommended to use tcpdump to diagnose the connection problem. Here's the results:
I issued the dhcp renewal commands in a terminal and monitored the output of tcpdump in a different terminal.
Code: root@xpsm1730:/home/lpallard# /etc/rc.d/rc.inet1 restart Polling for DHCP server on interface eth0: dhcpcd: version 5.2.2 starting dhcpcd: eth0: waiting for carrier dhcpcd: timed out
[Code]....
I believe the router is fine since all other machines are all using slackware 13.1 (one slackware 32bit and the other slackeware 64) and are all handshaking perfectly with the pfsense box.
I also believe the kernel on the faulty machine is OK since I am using the very same kernel(2.6.37) on the machine running slack64 and this machine is also handshaking perfectly with my pfsense box.
Please see http://www.linuxquestions.org/questi...get-ip-857683/ for more details.
I'm unable to get the "Uncomplicated Firewall" (UFW) to deny incoming pings. I've set it to deny incoming. Yet when I ping it, it responds.I'm using Debian 8.2 jessie KDE fresh/clean install with all updates.Below is the terminal output from a simple test (I've added an extra line feed between the commands for clarity):On a Debian 8.2 computer -- I install UFW, enable it, check its status (deny incoming), and get the Ethernet address:
Code: Select allroot@Computer:/home/user# apt-get install ufw .... (long output -- no errors or warnings) root@Computer:/home/user# ufw enable Firewall is active and enabled on system startup
[code]...
The Debian computer with UFW active is responding to the pings, when it should be ignoring them.Rebooting the Debian computer doesn't fix the problem.This setup is very simple. I'm using all UFW defaults.This is a new Debian 8.2 install (clean) with all updates.
