CentOS 5 Networking :: TCP Connections Stuck In FIN_WAIT2 State?
Jul 21, 2009
I'm developing a simple TCP server. 99% of incoming connections terminate correctly and the sockets disappear from netstat outputHowever, a few connections hang around for indefinitely in the FIN_WAIT2 state. Now, I know that the clients in this case are misbehaving by not sending a FIN,ACK to close the connection. However, regardless of client behaviour, the connections should only remain in this state for a maximum of 60 seconds (set globally by /proc/sys/net/ipv4/tcp_fin_timeout)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name Timer
tcp 0 0 10.0.0.12:2000 10.0.0.6:50990 FIN_WAIT2
[code]....
View 4 Replies
ADVERTISEMENT
Oct 26, 2010
Updated from 10.04 to 10.10 about 2 weeks ago and it worked just fine until today. After rebooting, it's stuck at "* Checking battery state" and nothing happens. I've tried with power cord inserted/out and without battery but with no success.
What I did before rebooting was that I tried to fix my iphone tethering problems with the following commands:
And also a system update. And here I am. Anyone knows what could have happened? I'm on a Acer Travelmate 8371.
View 8 Replies
View Related
May 17, 2011
I have a problem. I want interface to shut down once there is a link down state detected. If you set up eth0, and then plug cable out, eth0 stays up, just loosing the RUNNING flag. I want it to go down completely.
I've stumbled upon netplugd, and I've tried using it. It is a daemon that detects states in (link up) and out (link down) from kernel netlink interface. If I have this daemon up, it detects that link is down (eg. cable removed from interface), and it runs correctly /etc/netplug.d/netplug eth0 out. That script then runs ifdown eth0. And that's ok.
But very next moment netplugd runs /etc/netplug.d/netplug eth0 probe, and probe section of that script runs "ip link set eth0 up", which then turns interface back on!I've tried commenting out that line, but then new problem occurs - after the link is up, netplug cannot figure that out and bring back interface. Has anyone have this kind of problem, and any viable solution with netplug?
View 3 Replies
View Related
Apr 21, 2011
I just installed ubuntu on my m11x and am completely new to ubuntu. After installing the latest driver for the 335m I am stuck at checking battery state and there is no way I can get to the gui anymore. The only access I have to are the tty's and I don't know what to do. I have already tried finding a solution for a couple of hours, but cannot find any. Please help me solve this problem, I do not want to reinstall again.
View 2 Replies
View Related
Mar 30, 2011
I have a server that I can only access via SSH (it's located far away) and I would like to secure it by blocking all ports except the ones that I need (which are HTTP and SSH). I still want to be able to make outgoing connections to enable software updates and other things.This is my iptables -L -n :
Code:
Chain INPUT (policy ACCEPT)
target prot opt source destination
ACCEPT all -- 0.0.0.0/0 0.0.0.0/0 state RELATED,ESTABLISHED
DROP tcp -- 0.0.0.0/0 0.0.0.0/0 tcp dpts:1:21
DROP tcp -- 0.0.0.0/0 0.0.0.0/0 tcp dpts:23:79
DROP tcp -- 0.0.0.0/0 0.0.0.0/0 tcp dpts:81:65535
code....
In my opinion, this should block all incoming packets except the ones on port 80 and 22, but allow responses to outgoing connections. But a wget http://google.com does not work, it can't establish the connection.
Maybe this is not the best style for iptables rules, but I want to be absolutely sure to not accidently lock myself out from SSH, so I chose not to configure a "block-everything rule".
Does this configuration not enable incoming packets from connections initiated from inside?
View 3 Replies
View Related
Mar 30, 2009
I have a Intel WIreless 3945 card, wit ipw3945 installed on Dell E1505. I use NetworkManager to function and everything is ok. I can connect to wireless networks that have no security, or WEP security, but I just can't connect to networks with WPA or WPA2 securty. What can be my malfunction?
View 2 Replies
View Related
Jun 6, 2009
I'm a newbie in CentOS
I have a question
Does CentOS allow unlimited connections like Windows 2003 enterprise ?
View 1 Replies
View Related
Mar 26, 2009
I'm running Centos 2.6.18-53.1.13.el5 in a VM. Currently I'm studying for my RHCT.I'm trying to enable IP forwarding so I edited this line in the /etc/sysctl.conf:
# Controls IP packet forwarding
net.ipv4.ip_forward = 1
Then I ran sysctl -p
[code]....
View 4 Replies
View Related
Jun 5, 2010
This is on Centos 5.3 by the way..... I had a DVD and, well, just wanted to test some things.. I'm aware that this is not the latest release.
And, by the way (I just read an earlier post) I've added the username and password with which I'm trying to login (..hello?).
I'm at work and I don't have a copy of my Vsftpd configuration file (don't wanna edit this whole thing - actually there is a copy, below). I've used Slackware for years and never had any issues with Vsftpd - it just works. I am trying Centos because I need help with upgrading my mail server (qmail) and there's not much help for Slackware users. Also, I used to run RHEL3 and had vsftpd working fine (had to copy some file to /etc/pam.d) but it worked famously. I compiled all my web server stuff (just don't like these default things where I have no idea how things are integrated) and I'm hung up on - of all things - vsftpd.
Anyway....When I try to log into my ftp server I get "KERBEROS_V4 rejected as an authentication type" and only the anonymous account works. Any other logins produce an error (incorrect login - see ya' later). I have SE Linux and the firewall OFF. I don't recall the directive, precisely, but my vsftpd.conf file is set allow local users to connect. I installed with yum and it added some lines at the bottom (one was about a user list and the other was about PAM). I've got a chroot list and a user list although it's not clear to me precisely where the user list should be placed. I actually uninstalled the RPM and compiled, too. I've done everything but call an exorcist. And I've found tons of posts regarding this on the net and none of the fixes worked. Man - on slackware you type "make" and "make install" (I build it with tcp_wrappers) and you're off to the races.
Actually - I did upload the vsftpd.conf file to work (where I'm at, now). Minus lines that were commented out it looks like this:
Could it, possibly, be something about how I am adding the user, the shell type, etc? I know that in Redhat I used to type "/usr/sbin/useradd -d /home/someuser joe". I've done it that way and I've also done it like so: "useradd -d /home/schmoe -s /bin/bash schmoe".
View 2 Replies
View Related
Jan 24, 2011
i've got a select based application that wants to support a large number of mostly idle connections. the code is java and works on windows, suse enterprise linux, mac os x. it does not work on centos 5.5 (32-bit, 2.6.18 kernel, 1G of memory).
i've read and followed the directions in various articles about tuning linux for large numbers of connections (including the C10K problem), and gotten the number of sockets up to 3200.
these didn't make any apparent difference:
[URL]
on windows, i can get up to around 78,000.
on suse enterprise linux (a few years ago), i got up to 90,000. that's where i got bored and stopped.
on my mac laptop with os x (snow leopard), i got up to 10,500.
i have used ulimit -n 10240
my current goal is 10k sockets.
the test is that i'm opening one socket at a time until it fails. when it fails, many of the sockets which have already been opened also fail, in one giant cascade. sounds like a buffer / memory problem.
each group of 64 sockets gets a thread to manage select calls for them. thus i'm only using around 61 threads total when it fails.
View 3 Replies
View Related
Nov 13, 2010
i am running ubuntu 10.10 i tried installing nvidia drivers earlier. it said something about configuring X server as root. so i restarted my laptop and now i get stuck at the screen that says "checking state of battery" i know i can press ctrl alt F2. and i tried a few commands but nothing works.
View 4 Replies
View Related
Apr 7, 2009
After doing a clean base install of CentOS 5.3 on a test server, I started having problems with network disconnects. While SSH into the box, I would be intermittently disconnected from the SSH session. At first I thought it was my workstation, but after running ping tests from multiple machines on the network to the test box, it became apparent that the test box was dropping packets. When I lose my SSH connection, the pings also start timing out, for a few seconds, then start again. I increased the console logging and I can see when the ethernet port goes up and down when I physically disconnect the cable, but it does not go down or lose link when the disconnects occur. I changed the NIC ports being used and had the same result. I tried with a different switch with same result. I reinstalled CentOS 5.2 and the problem went away.
Upon further investigation, I found that CentOS 5.2 is using the e1000 driver for my NICs and CentOS 5.3 is using the e1000e driver. I resinstalled CentOS 5.3, and tried using the e1000 driver instead of the e1000e but it no longer detects my NIC's... apparently they moved the support from e1000 to e1000e. I have spent the last 3 days googleing for info on this problem and have not found a solution. Any recommendations? Should I submit a bug report?
[Code]...
View 12 Replies
View Related
Feb 7, 2011
have a problem with my network-manager in ubuntu 10.10.when I dial one of my vpn connections, my other vpn connections be disabled and I can't use them!I tried to restart network-manager and gnome-panel, but it does't seem to solve this problem.
View 1 Replies
View Related
Nov 4, 2009
I have a Centos 5.3 desktop that I have a development board connected via a serial connection to. The permissions on /dev/ttyS0 are currently set
crw--w---- 1 root root 4, 64 Nov 4 13:56 ttyS0
I have changed them so that they read
crw-rw---- 1 root uucp 4, 64 Nov 4 13:56 ttyS0
But every time I initialize the dev board the permissions switch back to the original state. I have found lots about preventing the permissions switching back upon reboot but that doesn't appear to be the problem.
View 5 Replies
View Related
Jul 14, 2011
I've got a CentOS 5.2 installation with broken package system. When I try to do "yum update" I get lots of "Missing Dependency" errors. How could I restore it to a usable state?
View 6 Replies
View Related
Jan 25, 2010
I just took delivery of a new computer, with CentOS 5 installed. It does not have a PS/2 keyboard socket, only USB. I get lots of strange "key stuck" problems - usually the shift key or the cursor keys (I guess I use them the most). Or sometimes key presses do not register. It makes it very difficult to use the machine.I have tried two keyboards (both of which work fine on other machines, or on this machine when booted with a live CD). The machine is an Intel i7. Are there any known problems with USB keyboard drivers in CentOS?
View 4 Replies
View Related
May 17, 2010
I am familiar with linux and have done some managing of servers before. I just tried to install centos 5.5 on a server, but after booting the screen was blank and my monitor said vga mode not supported. I found online to press ctrl+alt+F1, which allows me to see the rest of the boot commands. Then after I was able to see the boot commands it says enabling /etc/fstab swap, but never says ok or gets past that part.
View 2 Replies
View Related
Jun 8, 2011
with netstat i found a port with idle state, i google the meaning of IDLE:Idle, opened but not bound.What do you mean by bound?
View 2 Replies
View Related
Jan 28, 2010
I've got SAMBA and SWAP installed. I tried to localhost:901 and telnet localhost 901 And both connections tries were refused. I opened the port and even shut down iptables. What else do I need to do to use SWAT?
View 1 Replies
View Related
Nov 8, 2010
I've recently upgraded to 10.10 and am having the same problems as with all the previous versions. I have no eth0 connection at this moment, only lo. I have an Hp pavilion about three years old with a onboard ethernet, connected to router and wired broadband. The thing is, when I am in Windows, the internet will also drop, going into "local only." I can always get that back though by running the network diagnose. Another thing. How to I get into the network.state file to make changes?
View 2 Replies
View Related
Jul 12, 2011
Having a problem with tftpd on CentOS 5.5. I am only able to download files from the localhost address.
I have checked iptables (by adding rules and by turning the firewall completely off -- nothing changes) and enabled tftpd logging (by adding -v to the server_args section in xinetd.d/tftp) and I DO see the request coming in, but no response from the daemon. I've also tried upping the verbosity by adding multiple -vs and/or --verbosity 10 to the mix, but all I ever see is the initial request come in, nothing more. I've even tried running the daemon in foreground mode with -L but I get no verbosity then.
View 3 Replies
View Related
Jan 7, 2010
I've found a number of other entries about this, but they don't apply (the config files mentioned do not exist or whatever). My RHEL 4.x systems work differently, so the solutions I used there don't help either.In the course of some fussing around trying to get Virtual Box OSE to load, I apparently stomped on something so that when I log in now, I get the twm window manager--I was perfectly willing to have Gnome or whatever, but twm is a bit too light-weight for what I need.Anyone know where the magic config file or whatever lives in RHEL/CentOS 5.x?
View 3 Replies
View Related
Jan 20, 2009
the boot of my server is stuck at the message "GRUB"
I've allready tried :
chroot /mnt/sysimage
grub-install /dev/cciss/c0d0
but with no improvement. What else can I try?
View 6 Replies
View Related
Apr 1, 2011
I created a new Ubuntu 9.10 VM, and it installed fine. I then told it to upgrade, and was running it fullscreen and left it for several hours. Now the machine is stuck at full-screen, and Ctrl+Alt do not unstick it! I can see the mouse cursor, but cannot press Alt to bring up a menu or anything. I pressed Ctrl+Alt+F1 to get a console, and logged in as root and tried to kill-9 the process, (qemu-dm), but it won't die! Any ideas on how I can unstick the VM? I have lots of applications running and transferring data now, so I don't want to reboot the box for several days until the file transfers are done.
View 1 Replies
View Related
Aug 14, 2010
Upon boot of my system, I auto run a script that customizes the routing table (i.e., route -n). However, after the system is already booted up and has been running, if a network connection is dropped and re-added, like if one of the several network cables to the machine is temporarily disconnected or other, then the routing table is modified -- I assume by network manager. Presently, I need to manually re-run my script to re-customize the routing table.
I would like to have the script just run whenever there is any change to the connection state of interfaces. Anyone know how to tie in some kind of event trigger that can execute a script upon change in interfaces connection state (i.e., detect change in connection state and run the script)?
View 2 Replies
View Related
Jul 29, 2011
I am using RHEL6 on a server with 3 interfaces.
eth0 is connecting via a Cisco switch to the outer world and eth1 and eth2 to internal networks.
eth1 and eth2 are working without problems all the time, however I have the following incident with eth0: although it is RUNNING, and after it works for some time (like some hours), it then becomes incapable of sending/receiving traffic. Here are more details, after the interface exits the RUNNING state:
[root@mgt ~]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr E4:1F:13:6C:87:74
inet addr:9.162.91.26 Bcast:9.162.91.255 Mask:255.255.255.0
inet6 addr: fe80::e61f:13ff:fe6c:8774/64 Scope:Link
UP BROADCAST MULTICAST MTU:1500 Metric:1
[Code]....
I am wondering whether there this is because there is not traffic for some time and so the interface goes to idle. Is there any parameter I can tweak in RHEL6 to force the interface to remain RUNNING even in the absence of traffic (although I would be expecting that some 802.1 packets would be periodically flowing from the switch to the interface)?
View 1 Replies
View Related
May 28, 2009
I am writing a server which uses edge-triggered epoll. When the server calls accept(), it just extracts the first connection in the pending queue. If there are more connections that are waiting, can we make the server accept all the pending connections? I wrote a loop like the following:
do {
client_fd = accept(...);
/* Work with the client fd */
} while (client_fd != -1); [/code]
Doesn't seem to work. A related question, more a clarification, is as follows. My understanding was that a connect() at the client returns only when accept() in the server returns. Clearly, I was incorrect. Even with no accept() call in the server, my client was able to connect and send data. Am I getting this right?
View 2 Replies
View Related
Feb 7, 2010
I am stuck with 2.6.18-92.el5 kernel for a long time now. My centos doesn't want to load any other kernel. Seems like it doesn't read grub.conf file on reboot. Here is my grub.conf
[Code]....
View 14 Replies
View Related
May 16, 2011
I managed to install Centos x86 for the first time, together with VncServer, SSH access, mySQL and Apache. Everything worked greater until I logged in remotely to the server from my Windows 7 machine using RealVnc. I manage to connect successfully and able to use the GUI remotely. However, when I press logout from the menu, the screen gets stuck on the funky blue and white curly background. It does not close automatically.
If I close it using the 'X' button and reconnect, I get in to same stuck screen. My only workaround is to login with Putty and type 'service vncserver restart'. Then I am able to connect again. But same thing happens when I do a logout.
View 2 Replies
View Related
May 29, 2010
I am using Ubuntu Lucid Lynx (10.04)I have a wireless card in my PC that I only use very occasionally so I uncheck the Enable Wireless option on the network tool icon to stop the system attempting to connect to my wireless network. That works fine for that session, but the system does not seem to remember this setting and attempts to connect the wireless each time I log in! Is this a feature or a bug? and does anyone know a way of defaulting the wireless option to off
View 1 Replies
View Related
