Fedora :: Network Install Via Kickstart Using A - Created - Local Repository - F12 - F14
Jan 5, 2011
I am looking for a thorough document that explains:
1) Creating a local repo
2) Using kickstart to access that repo
3) Performing a network install using kickstart
Some background: I have several racks of servers that I need to install Fedora on. These servers CANNOT be placed on the internet; hence the need for the creation of a local repo on some other machine (which will be connected to the servers via a local network). I am not sure how to create a local repo, so that one of my questions.
I'd also like to automate as much of the install as possible and kickstart is the only thing I know of for that. I am no guru with kickstart, but I have used it before to successfully install Fedora Core 6 -- I am hoping there are no great changes with the current releases of Fedora (12-14)?
A local repo of Fedora Core 6 was created by someone (some time ago) on a workstation (running FC6). This is what I've used in the past to install FC6 on previous servers (via a kickstart CD). However, I dont have the documentation on how the repo was created or how the kickstart CD was created I've gleaned some ideas ok kickstart from the pieces I've read on web, but none of it has been specific to the latest releases of Fedora.
I have a kickstart script that attempts to create user "joeblow" with an encrypted password. The user is created okay, but the password does not seem to "take". After installation is complete, and the system is rebooted, this is the relevant portion of /etc/passwd:
Code:
This is the relevant portion of /etc/shadow:
Code:
Where there are two exclamation points, I would expect the encrypted password (as is the case for the root user, which is also created in the kickstart file).
The relevant line in my kickstart file looks like this:
Code:
That password string, which is 34 characters, is the word "password", encrypted with this command:
Code:
Also in my kickstart file is this line:
Code:
After the install is complete, I reboot and attempt to log in as "joeblow", but no dice. If I log in as root, and manually put the above encrypted password into /etc/shadow for jowblow, I can then log in as joeblow.
I am using RHEL5 and I did a createrepo -v --update /var/www/html/yum/base/5Server/x86_64/rhel-x86_64-server-5/getPackage/ and the output was
Scanning old repo data Indexed 10031 base nodes Killed
Now is that a error? I didn't press Ctrl-C or anything like that, it just finished that way. I have a local yum repository and I use reposync to sync up my local repository with the Red Hat network. Everytime I pull down the updates from Red Hat I used createrepo with no options but wanted to try the --update to see how much time it would save. Anyone have any ideas why it is it was "Killed"
I can't figure out how to copy a file off a custom RHEL 5 DVD that I created with kickstart after the installation completes and the %post runs. I don't want to run as "%post --no-chroot" because I have a script of system changes that needs to run in chroot. I just want to copy one RPM file into the /tmp so my script can install it. Can I have two %posts, one with chroot and one without? I can't seem to find any RHEL 5.5+ guides for kickstart that can explain to me how to make it work.
I wish to install debian on a number of boxes and have resolved on a network install. I'll first do a minimal install using the network install iso on a usb stick, then reboot and complete the installation using a local caching repo (apt-cacher) on the LAN. As a way of further minimizing bandwidth usage, I wonder if I could extract the .debs from a full installation cd and use them to populate the local repo?
Today I download a fedora beta iso file, then I try to install it following the steps I installed fedora 14. However, the install guide ask me to connect to network, if not , I had to exit installation. I just want to install fedora 15 from my local iso file. What's wrong?
I have been looking into setting up a local repository for updates etc as we have an increasing number of fedora clients/servers. Searching the web I found multiple how-to's on how to accomplish this little project. One thing that bothers me is that each how-to seems to refer to a "static URL" mirror for rsync to get the rpms.
I am organizing for the 14th march a Linux installation fest, and I would like to have a local server with all packets that I will need to complete the installation.BUT I would like to have these packets in a way that is usuable by yum, so like having a local repository server.You will agree with me that I cannot mirror the entire repository server (for disk space) but some packets yes, but I don't know how.
Well my local LUG has developed a cooperation with the university. Their major is music. So we would like to develop a local repository to help the updates to be faster. I checked the How to set up a local Yum Repository on Fedora 8. I guess it's same with 12. My questions are:
1. How much space do we need?
2. Since we might setup repository for another distro, is it a way to be done with other distro? The above address describes only how to make it in Fedora.
3. Since we'll have it installed and will be usefull for the university, can we make it available to the rest of the world? Maybe contact someone for this.
I'm confronted with a problem and haven't been able to find a solution using Google and the likes. I have created a netinstall image based on CentOS 5.6 with kickstart which works great while pointing to the HTTP / NFS repository in isolinux.cfg like this:Code:label publisherkernel vmlinuz append initrd=initrd.imgks=http://192.168.0.198/kickstart/ks.cfg cdncontentpublisherThe problem however is that the location of the kickstart file and repository/ies might vary and I don't want to create a separate ISO for each and every location.
It is not difficult to set up a regular local repository (rpms) and manage a fedora system with it using yum. I wonder if the process is same for deltaRPMS. Has anyone successfully set up a local deltaRPM repo or is this something reserved for online repos? How does one achieve this?
I have not used Fedora in a long while (since version 5 I think).I am moving back to it for my new home server but there is something that is driving me completely crazy. Can you no longer do HDD installs from a ISO?I whipped out my old kickstart files and worked off of them and all seams to be OK but every time it says it is unable to find the install source.Here are the first two lines of my KS file:
I am trying to use kickstart to create a master boot disk for all of my server installs. Unfortunately it seems to be missing some of the packages specified and I can't figure out why. Here is the packages section of my ks.cfg file
Note: all at symbols have been replaced with # because the forum filter seems to think they are urls.
I am preparing an installation for a network with small subnetworks across the country.As such, I am setting up a central repository, but would like the installation of some machines to turn themselves into mirrors at installation time (at least for their own rpms) to save bandwidth for the rest of the machines in their own sub network.Is it somehow possible to set up a local mirror at installation time?I was watching how anaconda is installing from a repository and saw that it erases each downloaded rpm as soon as it is installed, but thought maybe someone here would have an idea.
Using C++, I want to process sub-folders on my home folder sequentially each with a special naming format and containing some binary files in it:
Code: 1/ 2/ 3/ 4/ 5/ 6/ ...
Give above folders, I will process files in 1/ at first, 2/ at second, 3/ at third, and so on.
For some n/ folder, if I realize that n/ actually does not exist in local file system, I do not want to wait for it. Hence I will keep processing (n+1)/ folder, and so on.
However, when processing some (n+m)/ folder, previously not processed n/ folder may have been created on local file system. In this case, I do not want to miss processing it, but somehow detect its creation and process it. After processing n/ folder, I want to continue from (n+m+1)/.
I have the 5 Debian 5.02 DVD/ISOs, and would like to set up a local server from which to perform installations using a PXE bootloader and the local LAN. Any pointers to how I might accomplish this? All of the online resources I've found for network-based installs expect to do all downloading from a remote mirror, and I would prefer to limit my network traffic to the local LAN. Also, I want to install to multiple new hosts, and would like to avoid doing a lot of babysitting and swapping DVDs in and out of the drives (and some hosts are without DVD readers).
I have setup a deb mirror that is syncing the Ubuntu repositories to my local server. I am maintaining various versions of Ubuntu, from Hardy through to Lucid. I am trying to get a copy of just "Karmic" repo's to my local HDD. The problem is all the packages are under one folder /pool:
main multiverse restricted universe How would I sync all the packages for just Karmic and not the other distro's to my local HDD.
I am running Ubuntu 10.10 as my OS. I also have an .iso file of OpenSuse. It has a lot of packages on it. Can I add it as local repository so that instead of downloading the files from net, if they are available on DVD then I can use them.
I freshly installed debian lenny using the 5 DVD set that I downloaded from debian.org. I want to create a local repository for all packages that are available in DVD so that I do not insert the disc everytime I install a new software. I have searched various forums but not able to figure the right way to do it.
I used Acronis' Disk Director Suite 10.0 ["DDS"] to create 7 logical partitions of 23GB each, into one of which I asked Fedora 11 to install. Fedora 11 completely ignored me and created 2 primary partitions of its own: a 217.4GB, a 2GB and a 2GB "unallocated." I will likely delete this installation for a number of reasons. How can I force fedora to install into a 23GB logical partition that I created for just that purpose? If I can't use DDS-created partitions then why do I need DDS-created partitions?
The desktop computer stores most of our family's files and is mostly powered on, installed with Debian/openSUSE (KDE4).
We have 2 more portable (also Linux, KDE4) computers and frequently need to access files stored in the desktop. Currently file transferred is by usb thumbdrive which is troublesome.
The computers are all connected to a wireless router that provides access to the internet.
Reading/goolging and understand that there are many ways to share files to the extend that I am lost - Samba, NFS, SSH .....
My problem is to find out: What is the best way to access computers within the local network? "Best" here means:
- Secure
- Relative ease to setup
- Easy to use, file access with GUI (file manager?) as the entire family is using the computers.
I want to perform a netboot install of Ubuntu using files sourced from a local network FTP server, and I cannot find the necessary installer files anywhere.
Reasons include: * dead or flaky optical drive and no chance of replacing it * optical drive is CD but iso is DVD * several targets, so much faster to load repeatedly from local FTP than from an internet repo
What I can do with Fedora or CentOS is as follows: 1. On a second PC, download the relevant .iso and extract that to a directory (or mount it with -t iso9660 -o loop to make the contents visible without bothering to run the extract or using up more space). This makes all the files available but most particularly a ram fs and a mini linux installer, such as initrd.gz and vmlinuz
2. On that second PC, also set up the usual DHCP/PXELINUX/TFTP arrangements to cause the netbootable target PC to load those two files and run the installer. Also run FTP (or HTTP) on that PC
3. Netboot initrd.gz and vmlinuz (names vary) into the target PC. Answer the installer's questions, including choosing FTP (or HTTP) install method, giving the second PC's IP address, and an account and password, and path to the extracted iso. The rest is as usual.
I have not been able to do this with any Ubuntu netboot installer I have found so far. It either netboots, but then (seems pointless) wants a CD iso, or insists on pulling from a repo. Yes, I could create a local repo, but that seems to need a lot more work and space than setting up an iso on an ftp server. Is it simply the case that the Ubuntu family lacks this sort of netbootable installer or I have just been unlucky in searching for it.
I'd like a way to see all of the devices on my local network and what their local IP address is. I recall that I used wireshark to troubleshoot a similar problem a while back, but it doesn't seem to have a way to see all of the devices- only the traffic. (I'd like to do this without having to physically interface with my router if possible, and I am in an encrypted network if that matters)
I have installed a web server on my local network. Everything is well configured and web pages are shown correctly from Internet (outside the local network) using the domain or the public IP.The issue is if I try to see that web pages (using the domain or the public IP) from inside the local network. In that case the router config page (192.168.1.1) is shown instead of the web pages.From inside the local network I'm only able to see the web pages using the internal IP address (192.168.1.XX).
I use a local repository for ubuntu and when i use update manager it always shows gpg warning and shows that the information is out of date. Is there any way to suppress this warning. ( In Ubuntu 9.10 ) , so that update manager never compain about the local repository.
How to create a local APT repository for the intranet so that everyone in the local intranet can install directly from my local repository instead of communicating to the internet?
I would like to set up a local repository, that serves my LAN, and store it packages I found on the web (using pkgs.org for example) , can you please link me a guide regarding this ?
I have a Fedora 12 and OpenSuse DVD with me. It consists of lots of packages and hence to install these packages I don't have to download them. What I want to do is copy the package folder on my local machine and then add them as repository in Synaptic Packet Manager.
I've copied a number of deb files from my ubuntu Natty dvd to /home/bob/Natty-DVD/pool. And I ran: sudo dpkg-scanpackages . /dev/null | gzip -c9 > Packages.gz Which created the nesc. Packages.gz file. Then I edited my /etc/apt/sources.list and added the line: deb file:///home/bob/Natty-DVD ./ and did a sudo apt-get update But, the list of repos doesn't include my /home/bob/Natty-DVD. This is on a fresh install of natty.
