Server :: An Interesting Challenge Involving Diskless Booting A PC?
May 13, 2010An interesting challenge involving diskless booting a PC
View 1 RepliesAn interesting challenge involving diskless booting a PC
View 1 RepliesWhen I run:Quote:mount -t davfs http://xxxx/webdav /home/USER1I get:Quote:Please enter the username to authenticate with serverhttp://xxxx/webdav or hit enter for none.
Username: USER1
Please enter the password to authenticate user USER1 with server
http://xxxx/webdav or hit enter for none.
[code]....
I just noticed the results of the Honeynet Project's Challenge 7: Forensic Analysis of a Compromised Server have finally been posted today. Just got done reading one of the submissions and it's pretty good if anyone is interested in how to analyze a Linux incident involving evidence from memory and the file system.
View 2 Replies View RelatedI'm attempting to set up a server for performing diskless boots here. I'm using Gentoo 2008, and have followed the instructions here:
[URL]
I've gotten to the end of the basic DHCP setup, and I can not get the server to issue addresses. I've confirmed dhcpd loads up OK, and is running in top.
Here's the contents of /etc/dhcp/dhcpd:
Code:
# Disable dynamic DNS
ddns-update-style none;
# Assume one default gateway for IP traffic will do
[Code]....
I currently have a debian lenny high performance cluster running seamlessly using dhcp and tftpd-hpa. All nodes are completely diskless, and i would like to keep it that way. I now would like to have a client boot over local into a full KDE desktop.
Here's the cache.
1. I would like the client to also be a node. I would like it to handle processing as the other nodes do, but this one with a full desktop, with ALL data and privileges accessible.
2. The client will have its own workstation graphics. It will need to have that driver loaded without slowing the system.
3. Client will need to be able to use its own usb, bd-rom, dvd, audio, etc. even though it will be booting into the HPC.
4. Wireless. I cannot run 70 ft of lan across an apartment. not happening.
The idea is a media center applicable "powerhouse" with high-end power and graphics for easy media editing, 3d creation, and gaming. Will also be used for webcache over all local, file sharing, and other server applications.
HPC specs:
-headnode: hp compaq p4-ht 3.2 ghz, 4x500gb, 4gb mem, gigabit ethernet, wireless-N
-nodes: 4x p4-ht 2.2 ghz, 2gb mem, no-name brand slim desktops; 2x panasonic toughbooks p4-mobile 1.8 ghz, 1.5gb mem; 3x gateway p3 unknown specs; and 1 hp 530 notebook 1.6ghz core2 duo, 1.5gb mem
-client: hp compaq core2 duo 2.13 ghz, 4gb mem, wireless n, bd-rom, 2x dvd-rom, usb 3.0 card, bluetooth, 1gb AGP nvidia geforce 6800 dual-dvi, no hd.
I need to know how to set up this PXE.
My problem is that I can't setup a diskless environment with CentOS 5.4 (server) and CentOS 3.9 (hosts). On the host, I've recompiled the CentOS 3.9 kernel, with these mods:
- added kernel level IP autoconfiguration;
- added NFS file system support;
- added Root file system on NFS;
- added e1000 driver (the host has a gigabit network card);
Then, I have followed the guide at this link: [URL]
But in my dhcpd.conf I'he put other parameters:
ddns-update-style interim;
allow bootp;
allow booting;
subnet 192.168.195.0 netmask 255.255.255.0 {
range192.168.195.70 192.168.195.90;
option routers192.168.195.254;
option subnet-mask255.255.255.0;
option domain-name"lnf.infn.it";
option domain-name-servers192.168.195.254;
option time-offset-18000; # Eastern Standard Time
default-lease-time21600;
max-lease-time43200;
filename"linux-install/pxelinux.0";
next-server192.168.195.254;
}
host dante79 {
option host-name"dante79";
hardware ethernet00:20:38:01:C3:7C;
fixed-address192.168.195.79;
}
I have a PXE server which installed DHCP TFTP AND NFS service. and set them up followed by [URL].
The boot process hanged here:
ipconfig: eth0: SIOCGIFUNDEX: No such device
ipconfig: no devices to configure
[-n eho]
. /tmp/net-eth0.conf
/init: line 3: can't open /tmp/net-eth0.conf
...
What I did try to work around is,
1. Set portfast to all the ports connected the PXE server and client.
2. Edit pxelinux.cfg/default, try different setting in "ip=" (I did not find any menu in the internet which explain to to configure this file).
I am having a few problems with a red hat box involving memory usage. I have 64Gb memory and 'top' tells me I'm using 60Gb of it, but if I add up all the '%MEM' figures I get no more than 20%. Where is the other 80% ?
We have an ORACLE instance that is using shared memory but this is ceilinged at 45Gb. That means there is about 15Gb unnaccounted for . What utilities can I use on red hat to ascertain memory usage other than 'top' ? Any better ones, more detailed, looking at shared memory etc and swap ?
I have used the words 'Open Challenge'.I have attached a spreadsheet 'Sample file for filter.ods'.TASK: In a presentation, I would like to show what are the Topics of India on Day-1 , Day-2 , Day-3 , Day-4. That means I am going to Filter India and (Day-1, Day-2, Day-3, Day-4).
View 7 Replies View Relatedinteresting commands for the command line so I thought I'd ask people who've used Linux a lot. The only one I know is:jp2a (JPG to ASCII), a small program which converts JPG images to ASCII. It's an amazingly fun program to use, and everybody think it's awesome to be able to view images in ASCII. If you want to view a photo which isn't JPG you can simply use the convert program and pipe it in.
View 12 Replies View RelatedI'm looking for a screensaver application which will show random photos from my collection, but rather than just changing between images, I'd like something which shows them in a cool way, for example, on my brother's Mac, the screensaver displays photos, but as if they're being dropped onto a surface, i.e. not always with the same orientation.
See here: [URL]
This is a list of softwares that I use regularly. Some are OSS and free. Some are not. Just figured I'd share it, even though its primarily for me.
View 2 Replies View RelatedI'm trying to configure NFS sharing behind a firewall, I got it to work and all but I was caught by something that (to me anyways) seems odd.I've been able to mount the export on another computer and am transferring files over as we speak, but I'm just interested in knowing why the RELATED,ESTABLISHED rule seems to be catching almost all the traffic coming from the other node. Any ideas? Should I be concerned that my firewall isn't protecting anything or something?
View 1 Replies View RelatedI have a Toshiba laptop that seems to be freezing on extended operations involving the file system/hard drive. (that's kind of a guess on my part). The system has frozen during the last day or two when I:
- try a large file operation with Nautilus
- try exporting a large video file from Kdenlive
- try formatting the disk for a new install of Maverick (I decided that with the new version almost out, I might try upgrading in case I broke my system)
What I see is the screen freezing and the mouse does not move. As I said, I tried installing the Maverick RC from the Alternate CD (I wanted to do a minimal install). The install hung on the partitioning step and the caps-lock key started blinking (though I don't get this in other cases). I have booted into my Lucid live-cd and a Puppy live-cd to try Gparted, with the same results - the system freezes (but the caps-lock key doesn't blink). Puppy HAS performed better generally, for example I was able to complete my large file operation (backing things up!) but it is freezing on the partitioning step.
I'm trying to install centos 5.3 as part of AsteriskNow 1.5 on a SuperMicro server with main board type X5DA8 The anaconda fails with the following text visible:
[Code]....
Other terminals (F3 and F4) don't appear to have any interesting information on why it barfed. I tried it with a text install too, much the same result. I get this error with other Centos 5.n based Asterisk and GUI bundled mini distors too, specifically TrixBox
we usually have to access a batch of computers that is located on campus. This is straightforward enough on Windows, the OS that the computer support center cares about, but I'm interested to see how I would handle this in Debian. The process for Windows is:
1) Run the authentication program, which runs in the background.
2) Use Exceed on Demand (or Putty, but I can't really use that for my classes ), and log into the server. From there, it lets you log in to whatever computer you want. When logging on with Exceed on Demand, Solaris environment is then launched.
Now, I know there's not a lot of technical information here, and I apologize, but I know next to nothing about networking. What I do know is that I'm remotely accessing a secure computer, and then I ssh over to another computer (when using Putty, which was only command-line). I was wondering how easy or hard this would be to do through Debian, including the fact that I need to have the authentication program running.
I have a mythfrontend box that boots diskless - everything works fine. However, when attempting to upgrade the kernel - grub complains about not finding a device, and the upgrade fails.How do I fix this? Other packages upgrade with no problem.
Code:
The following partially installed packages will be configured:
grub-pc linux-generic linux-image-2.6.32-23-generic linux-image-2.6.32-24-generic linux-image-generic
[code]....
I have followed everything in the fedora 14 documenation about setting up a diskless pxe boot and I can get past initialization. The kernel and initrd load fine from the tftp server. The kernel starts, then hangs after probing USB devices. The messages show that initrd does not have support for nfsroot. I think I need to better understand how to add this in when I run dracut to build the initramfs image.
View 1 Replies View RelatedI made a diskless image against Fedora15, during the boot I found it displayed the following error message and went into emergency mode.
The error message:
Starting Relabel all filesystems, if necessary ^[[1;31maborted^[[0m because a dependency failed.^M
[ 107.607155] systemd[1]: Job fedora-autorelabel.service/start failed with result 'dependency'.^M
Starting Mark the need to relabel after reboot ^[[1;31maborted^[[0m because a dependency failed.^M
[code]...
That is, the thin clients will have to get their OS image via TFTP, and their IP address via DHCP on a *wireless card* - is this possible?What spec should I ask for/search for when buying the WiFi card?
View 1 Replies View RelatedI'm trying to set-up a diskless environment on my home network, based on a F12 server and one FC14 client and I have a problem that I think is related to NFS, but googling around I cannot find any solution.
Basicly the client boots correctly, gdm fires up showing the login window, but the system doesn't let me in, if I switch to console (CTRL+ALT+F2) I cannot login with both standard and super user getting a "permission denied" error.
I tryed to mount the exported fs on another computer and ls -l on the mounted dir shows both user and group set to nobody.
If i mount the exported dir on server side (i.e. mount -t nfs localhost:/...) permissions user and group shows correctly
In short we are booting Centos 5.4 over PXE to a bunch of diskless clients. Once they are booted - we can login (as local root account) and RDP to windows machines using rdesktop as we require.
The next step of the project is to get user authentication to the Windows Domain controller working for the PXE image.
To do so - we continued with our physical install of Centos 5.4 (used to create the pxe image with rsync as per the wiki page for diskless clients) by following through this page. AD auth works perfect on this box (it has a local HDD install of centos obviously).
Once we rsync'd the changes over to the pxeboot location - and rebooted one of the diskless pxe clients - we get issues.
The issue is that winbind seems to start - however the file "/etc/samba/secrets.tdb" cant be read. We tried removing this file that the PXE clients use and recreating it using
touch /etc/samba/secrets.tdb
I am currently working on a Diskless Fedora10 based PXE image that should auto login and then launch the ICA client without launching the desktop (so normal autostart will not cut it)I have got the image up and running, the clients boot correctly, autologin etc, but still launch KDE, and I was hoping replace the desktop with the ICA Client.
View 1 Replies View Relatedhow to to setup a PXE boot environment. The server contains a LUbuntu 10.10-installation the client is Ubuntu 10.10 Server (I didn't mix up those two...) So, everything is going fine; the client gets an ip-adres, and boots over PXE. The login-prompt is shown. Problem: I can't login. It seems my users are deleted (better not copied). Of course the system (client) works fine when I boot the system from it's harddrive. The only thing I can guess, there going something wrong with:
[Code]....
I have one router, a linksys. It allows wireless and wired connections, as is normal. I have two XP machinesby wire to the router and three linux machines connected wirelessly. The XP machines both have IP addresses beginning with 192.168. while my three linux machines have IP addresses that all begin with 172. None of the machines is connected with a static IP address. All are automatic DHCP.I am told that the above scenario makes no sense. However, such is what I have so, I trust, the theory and the fact do not gel. I would not care except that I cannot see - using the nautilus network servers program, all of the XP computers with some of my linux boxes.
View 9 Replies View RelatedI am trying to resolve an ongoing fight involving /dev/video node assignment between 2 TV Tuner cards and a USB video camera.
Each time I reboot the three devices seem to shuffle their /dev/video assignment. i.e. what was /dev/video0 (USB camera) after a reboot magically becomes /dev/video1.
This causes all kinds of grief between my MythBackend and ZoneMinder who expect the devices to have static assignments.
how to best solve this? I did a brief search and one solution appears to be to statically assign the devices.
some basic programming/scripting/etc. ability but I am not able to do what I now need to do. I would like to have a link on a website that when a user clicked on that link, one script or another would execute based on what operating system that person is using. To wit: If a Mac User clicked on it, it would run a terminal script and would set up a served printer for him; if a PC user clicked on it, it would run a script to set up the served printer on his machine, you get the drift. The printer is being served on a Windows Server 2003 machine, the users are NOT domain users, the print server is also hosting the files, scripts, webpage.
View 11 Replies View RelatedPerty self explanitory, I'm trying to boot up ubunto server edition 10.10 from a USB stick (pen drive) and it works fine until step 3 when it says "Your installation CD-ROM couldn't be mounted. This probably means that the CD-ROM was not put in the drive. If so you can insert it and try again
Try again to mount the CD-ROM? <Yes> <No>"
I don't have an installation cd-rom! I put the .iso file onto a memory stick, inserted it into my old computer (which i will turn into a server later) , restarted it and got past the first two slides and got this error? Do I have to put the iso on a cd-rom?
I have been trying to accomplish this feat for about a week now, and I always seem to get a boot up error.
Here is how I call WinPe from my default file
Code:
We have done an installation on a blade server (HP) for RHEL4.5.
The server have an HP smart array 2200 controller and during installation it can see the controller and disks just fine.
The installation completes up to the point where the system needs to reboot , however after the reboot it does not start up again and get stuck at the message: "Trying to boot from drive c:"
We used the rescue option to mount the installation and cant find nythign wrong, even with the grub setup (except that it points to hd96 as its primary disk). we tried changing this to hd0 as well but still no luck.
After about 3 installations we are still not having any luck.