So my server running FC4 died last night and I decided to go ahead with the long-awaited upgrade to FC10 while I was rebuilding my server. I use my server for a number of things including, but not limited to: router, firewall, web server, mail server... I have a typical configuration process that I have followed since before fedora, and it has worked well for me up through FC4, but my usual config process doesn't work on FC10.
First of all, I don't want SELinux running, I didn't see an option to not install it during the FC10 setup, so how to I properly disable or uninstall it afterwards? Second, I was unable to even configure the server to route traffic from my internal network to the web, here's the process I usually go through for this:
[Code]...
I don't have a static IP from my ISP, so I'm not quite sure how to add the default route. I think I need to do something like "route add default gw xxx.xxx.xxx.xxx eth1" where eth1 is my external NIC, correct?
What else do I need to do to route traffic? I noticed that I wasn't even able to ping my server from the internal network even though they are on the same subnet, my server's internal NIC has an address of 192.168.7.1 and my computer on the network has an address of 192.168.7.2. If I can get this going to I have internet access again at the very least, I can move forward with the web server, email, etc.
I recently built my second general-purpose server, and recently installed fedora core 10 on it. The first thing I attempted to set up after installation was the network - and that's where it's gone wrong When editing a network device using the graphical system-config-network utility, I find that the subnet mask is being automatically changed to match the default gateway address every time I attempt to modify any of its settings (or sometimes even when I cancel the changes). This also means that I cannot set the subnet mask, as it simply won't accept my setting for it. I seem to be able to get around this glitch by setting the subnet mask using the shell version of the same utility, but that doesn't solve my network issue.
Even when I use the shell utility to fix the subnet mask, I'm unable to ping other computers or routers on the network even when ifconfig indicates that the desired ip address has been taken, and other computers on the network are also unable to see the server. I'm using a wired connection and a static IP address on a network with no DHCP.
just as the title states i need to upgrade my os from fc4 to fc10 first let me say i have no physical access to the box. all i have is root ssh access to it. So i need to know the best way of upgrading that will be safe. before i do the upgrade the server will be backed up so i will not loose any files. if the upgrade fails the data center will reload the os with what i need on it. but i would rather not have to format.
I want to backup my apache website using tar. I want to make sure all the permissions and such stay the same so that if I restore a website, it's ready to go exactly as it had been.I am using the following command and would like to know if there is anything else I should neeed to do?
tar --xattrs cvzf backup.tgz /path-to-webserver-files
Now my other question is since tar doesn't store absolute paths, when I run the restore I have to be in the root directory? In another words if I run the command in a user directory will tar do something retarded like encode the paths as ../path-to-webserver or will it just be path-to-webserver so that when I run the restore from the / folder it will automatically go into the right place?
I recently setup a file sharing server with FC10 running Samba...
I got it running perfectly and it ran great for a few weeks. I was working on the machine today and tried to reboot it and noticed that it was stuck on FC10's new boot screen. I pressed the esc key and saw that the boot process was stuck on the smb service...
I don't get any errors and the machine doesn't freeze up or anything, it just hangs on the smb service and won't continue with the boot...
Is there anyway that I can get the system to skip the smb service so that I can get into the system? I don't know what's changed or what's causing this sudden hang in the system at boot.
I tried the ubuntu live CD to attempt to go in and edit the /etc/rc5.d scripts to try and skip the smb service start but I am not sure where to look for it...
I made a big mistake on my FC10 machine-while trying to resolve a way to install vlc- I wrote the following command:"yum erase rpm lib".After restarting the machine fedora did not load... (might be funny- I laughed...).Do you know of a way top recover from this state? (I have the liveCD disk),
I want to use tab networking in my kvm with routing.Can any one guide me how i can do it. i have been reading different guides over the net but not understand any one clearly.I have read this[URL].. One problem is this all my server are remote and no gui is running.I am able to install kvm with ssh console with -nographic and -x "console=ttyS0" option now i want to change from bridging to tap networking with routing.And i have live ip on kvm guest/Virtual machine.
Here's the deal: I had a nice little fileserver running under 2.6.27.21-170.2.56.fc10.x86_64.
3 disks in raid 5 ext4fs, then I thought..."hey I'm a greedy bastard..I want another drive!!
So I get it..do a normal mdadm --grow...after around 1100 minutes .. FINISHED!..whee..happy...
I decide to do a upgrade to 2.6.29.2-52.fc10.x86_64 to get the fix for growing the ext4..
reboot...
Code: md: bind<sda3> md: sdc3 has same UUID but different superblock to sda3 md: sdc3 has different UUID to sda3 md: export_rdev(sdc3) md: sdd3 has same UUID but different superblock to sda3
I have a lab of 10 computers with ips from 192.168.1.120 to 192.168.1.130 the server's Ip is 192.168.1.116When I am on client computer I type the server's Ip address on the browser and it works. All i want is that instead of entering my servers Ip I could just enter an address like: example.lan
I'm having a problem getting BIND to start after doing the FC10->FC11 upgrade. Others have been reporting some bad symbolic links, but this isn't that problem. I've searched the forums here, but I don't see anything like this anywhere.
I have encounter strange situation I have set all my network configuration via /etc/sysconfig/network-scripts/ifcfg-eth0 ,eth0.2 ,eth1 etc .... (CentOs 5.5). I have used only once Gnome network manage settings to add routing since then all my network config change although all files mention above showing the same config I have set initially and same for Gnome network Manager.
I'm relatively new to Linux, but willing to take a plunge and climb the steep learning curve! I've recently installed Fedora 10 on a spare box I'd like to have running in the lounge by the TV as a TV recorder. Before starting out with MythTV, I would like to get the box running the basics... and hit no major snags with display / audio drivers. I've finally managed to connect to the box via VNC from my macbook over the wireless network. The Apple standard client is slow over wifi, so have started using 'Chicken of the VNC' as the client. At the server end, I intalled vnc-server via yum. I've also tweaked things in ~/.vnc/xstartup so that my default window manager over vnc is gnome.
I'm running Ubuntu Server 11.04 with OpenSSH, trying to create an ssh tunnel (for web traffic) to it from my (also Ubuntu) laptop. This is the command I'm using to create the tunnel:
Code: ssh -ND localhost:8080 george@192.168.1.20 I had it all working on a virtual machine.. which was deleted What settings/lines do I need to change/add from the default OpenSSH config files to get tunnelling to work? I've Googled and AllowTcpForwarding is set to yes, as is X11Forwarding.. but it still doesn't work. Chrome can connect to the server, but says the connection was closed before any data was sent.
have a sample kernel .config, so I can compile a small custom kernel on my vps? I tried to make it alone,but I removed stuff,so that my VPS won't start anymore. Virtualization is KVM, running on a Core 2 Duo.
I've read through these pages [URL] but can't seem to get Samba server working. I am using Gnome and I go to System - Administration - Services
In there I can see smb and nmb. Both are enabled, but both have a status of unknown. I can't seem to start or stop either of them. I can just enable or disable them. I've tried using the gui, as well as command line.
I'd really like to get this going as I'd like to copy files from my XP Pro box to my Linux box (F9). I have WinSCP, but am having permission issues and I can only copy to a folder on my desktop on the Linux box (I'd like to get this fixed too, but that's a different thread at a later time)
I have a computer running Fedora 11. I wanted to use samba in my office. So I configured the /etc/samba/smb.conf ,but it did not work well : Windows can find samba,but when I tap in username and password,samba request them again , and the same thing continues in a loop. So I copied another computer's smb.conf (which is also a Fedora 11,samba runs well in this one).I copied this file to /etc/samba/ ,then samba can't satart up . I read from /var/log/samba/log.smbd.log : Unable to open configuration file "/etc/samba/smb.conf": Permission denied Even if I use: chmod 777 /etc/samba/smb.conf
why I can't login samba from Windows:
[global] workgroup = MYGROUP server string = Samba Server Version %v ; netbios name = MYSERVER ; interfaces = lo eth0 192.168.1.99/24
I am installing VMware-server-2.0.2-203138.i386 on FC 12
Have an error when executing vmware-config.pl initially. When I try again it has the message:-
start of messages;
The following VMware kernel modules have been found on your system that were not installed by the VMware Installer. Please remove them then run this installer again.
vmmemctl vmci vmxnet vmhgfs vmblock
I.e. - 'rm /lib/modules/2.6.31.12-174.2.3.fc12.i686/misc/<ModuleName>.{o,ko}'
Execution aborted.
End of message
I tried remove them using modprobe e.g.
modprobe -r -v vmemctl the response was "WARNING: All config files need .conf: /etc/modprobe.d/anaconda, it will be ignored in a future release."
Subsequently I tried vmware-config.pl, but have the same error.
I just done a brand new install of fed12 and did all the yum updates. Apache seems to start ok and I always liked the http config tool but it won't run on Fedora 12. I downloaded and installed system-config-httpd.noarch 5:1.4.6-1.fc12 and it all went fine but when I try to start it I get the usual box asking for my root password, I type it in and press enter, the box disappears and then....nothing. If I run system-config-httpd in a terminal I get the same box asking for root passowrd but when I enter it I get a long scroll of text which ends with:
line 4: 2137 Aborted (core dumped) /usr/bin/python /usr/share/system-config-httpd/ApacheConf.py
I don't know what causes this. Is there any way to get the http config tool working?
I've got a situation where I would normally use NFS, but cannot. So in it's place I need to use a samba share (even though it's linux -> linux). I need it completely open and world writeable as if I had done an rw in an NFS export. I thought I had it as when logged in as a user I can edit delete etc. however the apache user seems to be struggling with creating files.
Here's my smb.conf as it stands. Code: [global] workgroup = WGRP server string = Samba Server Version %v security = user passdb backend = tdbsam unix extensions = no
[sites] comment = Sites Directory path = /var/www/sites public = yes guest ok = no write list = siteusr writeable = yes create mask = 0777 browseable = no security mask = 0777 directory security mask = 0777 note that the drive is mapped as user siteusr
I'm trying to config a customer's samba server to use a MS 2003 SBS for auth. The samba is for file sharing only config'ed with disk quota's. Currently the samba is config'ed for auth = Users. But I want a central point for user auth, and since they already have the sbs in place, I'd like to use that.
Configure the HTTP server via system-config-httpd brings up some nice gui.I start a console, login to root and start it with: system-config-httpdGoto tab Virtual Hosts, then addI name my Virtual Hosts "sea", set its document root to: /home/sea/webs/sea.hs/www where index.php is located for that 'local test server'.Host information is on Name based Virtual Host.But what do i have to enter for IP Adress and Host Name? Neither of these accept 127.0.0.1 nor localhost Further, when i try to exit, it wont let me close the app.Its not saved, hence i can even delete the default virtual hosts and its there again after restarting the app Even with adding sea /home/sea/webs/sea.hs/www or sea localhost.localdomain localhost to /etc/hosts doesnt change anything
routing tables using "ip route 2" I have a server(server1 from now on) with eth0(internet connection) and eth1(lan connection). eth1 have 2 more alias devices = eth1:1 and eth1:2 On my server 1 the eth0=public ip, on eth1=192.168.10.1/24, eth1:1=192.168.20.1/24 and eth1:2=192.168.30.1/24 server1 is running squid and iptables to all 3 networks (eth1, eth1:1 and eth1:2) All of the clients have access to internet. Now what i want to do is add 3 more networks via a router(linuxBox = server2 from now on) connected to server1 in eth1. so the network will be like this:
Code:
server1 eth0=internet eth1=192.168.10.1/24 (connected to server2 and other clients)
I have an Ubuntu 9.10 server installation which have been working flawlessly for some months. The server runs bridged networking, because of some VMs that runs on it. But, a couple of weeks ago the network connectivity have started to disappear now and then (usually once a day or so). Running "sudo /etc/init.d/networking restart" always kicks it back to life.
After a bit of debugging I noticed that when OK the routing table looked like this: Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 192.168.1.0 * 255.255.255.0 U 0 0 0 br0 192.168.122.0 * 255.255.255.0 U 0 0 0 virbr0 default 192.168.1.1 0.0.0.0 UG 100 0 0 br0
While when in the non-working state it looked like this: Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 192.168.1.0 * 255.255.255.0 U 0 0 0 br0 192.168.1.0 * 255.255.255.0 U 0 0 0 eth0 192.168.122.0 * 255.255.255.0 U 0 0 0 virbr0 default 192.168.1.1 0.0.0.0 UG 100 0 0 eth0 default 192.168.1.1 0.0.0.0 UG 100 0 0 br0
Deleting the two routes to eth0 restores the networking again. Why these "erronous" routes gets added, and what adds them? How to further debug this?
I have one external ip address and a few domains. Would it be possible to have each domain on a internal domain and the box that sits on that external ip route to the internal. This would be for a number of server (mail, apache, imap, pop3, https ) So for example:
Some visit domain1.com ->external ip -> 192.168.10.100 Someone else visit domain2.com ->external ip -> 192.168.10.101 and so one with a number of domains
I have an environment of roughly 30 machines that all have ssmtp installed with identical config files. I also have logwatch installed on all of them, and it runs nightly as it is supposed to. The problem is that any given night, a random number of machines do not send out the resulting email from logwatch but instead dump it to ~/dead.letter. The number of failures changes every night, but most of the time it is between 20 and 30 of my servers.
Just curious to see what everyone's opinion on using routing vs. bridging for openVPN. I'm installing openVPN on a linux box that I'm using as a router. What I was wondering was your opinions on which one of these two options to use.
