Newsgroups: gmane.linux.debian.user Date: Wed, 4 Jan 2006 19:51:17 +0100
Newsgroups: gmane.linux.debian.user Date: Wed, 4 Jan 2006 19:51:17 +0100
> I've just installed another Debian Stable and I want to know why on earth the Debian maintainers believe it acceptable to just go ahead and install Exim4 without asking, forcing me to go and uninstall it afterwards....
Because exim4 is one of the best MTA's and historical it IS the defsult.
Each UNIX-like System needs a MTA.
> I'd like there to be a minimal install option but I can't see it, the closest I've found is the manually selecting packages and then just not installing any. I'm very used to Gentoo which gives full control over such stuff and I've been making the case for Debian at work, I'd really love to be able to control this a bit better.
Then create your customized minimal-system
apt-get install debian-cd
Michelle Konzack
Debian-cd is the official tool for building Debian CD set since the potato release. It was formerly called YACS (for Yet Another CD Script).
Its goal is to facilitate the creation of customized Debian CD set.
Posted by michael on Mon Dec 08,
hsoom writes "Debian Planet is reporting that unofficial sarge-based ISOs using the Anaconda installer can be downloaded from here.
The features developed so far include '…changed the code that installs software to use APT instead of RPM, removed Red Hat-specific configuration hooks, and written a new tool called picax that builds Anaconda-based installation CDs from a Debian repository'. However there are features that are not yet working and it is not recommended for use in a production environment."
Here's the link to building anaconda-based debian ISO images. [progeny.com] http://platform.progeny.com/anaconda/how-to.html
Finally a quick, easy way to remaster debian to hand out to friends.
Not to excited (Score:4, Informative) by killmuji on Monday December 08
Before getting too enthusiatic about this, please do remember to read the errata [progeny.com] before downloading the iso images. Lots of work still needs to be done, but this is a step in the right direction.
Anaconda won't be Debian's default installer — the next version of Debian will use the new Debian Installer [debian.org], which supports multiple UIs and all the Debian platforms.
Anaconda has been ported to Debian by Progeny, mainly because Progeny supports both Red Hat and Debian and they want to use the same installer for both distros.
Oh, and yes, Anaconda can be run in text mode, but it doesn't currently work [progeny.com] in the Progeny port.
Hopefully this means we have Kickstart too.
Debian has been needing kickstart-like functionality for a while. (No, FAI is not the answer, it works in a somewhat different manner, and its a royal pain to set up to bootstrap unstable systems from a host running stable).
> One of the main 'comments' I get when I recommend Debian GNU/Linux to > people, is 'Debian is difficult to install'
I think it can be argued that the Debian installer asks many questions that may not be easy to answer for a Linux newbie.
But, as you say, there is hope: I remember someone saying, a few years ago, that a RedHat had formatted their drives without clearly mentioning that it would be destructive (oops!). Today, Mandrake can be installed after just a few minutes worth of clicking "OK". It generally makes the right choices for the user, clearly shows what partitions will be created, and warns if it's about to blank an existing windows partition. If it finds some unsupported hardware, it mentions what it knows about it, so that the user can simply ask their local guru for help.
> I think it can be argued that the Debian installer asks many questions that > may not be easy to answer for a Linux newbie.
That's true, but the point is that most of those questions are unecessary. Every PCI device on my system has a number, and the Linux kernel comes with a map of device number to driver names.
I'm not a Linux newbie, but see it as a waste of time to answer silly questions about the modules I'd like to load for every bit of hardware in my system when my hardware is designed explicitly to avoid the need to do so.
T
Date: Thu, 6 Sep 2001 11:25:14 +0200 Cc: debian-user@lists.debian.org
> After using linux for some years, I now know which applications
> I like best for each task (text editing, graphics, www browsing,...).
> for the same reason, I also realize that I always end up with a lot
> of unneded stuff on my HD, unless I do spend a lot of time picking
> packages by hand, with trial and error.
>
> Now, let's assume that I have successfully installed the bare, bare
> debian system on a PC. At this point, **without net access**, what is
> the best (script based) way to:
>
> 1) install X11 with only the server for the card I actually
> have
>
> 2) declare the *single applications* (not packages or packages
> group)
> I need, and let the system sort all the dependencies out and put
> on the PC the *minimum* set of debs needed (picking them from
> *all*
> the CDs)?
>
> In other words, I would like to be able to say through a script:
>
> **I** want to use {gcc, perl, mutt, apache, postfix,
> cups and not lpd,
> pppd, iptables, ssh
> blackbox and no other file managers,
> emacs without X support, bash, fetchmail,
> gimp and gimp only, no other GNOME things
> konqueror and no other kde things
> ....add your favourite apps here
>
> }
>
> **YOU** (bare debian system) sort out all the dependencies and
> pick only what is needed from the CD(s) while I go to lunch
>
> keep in mind that I don't need some automatic multiple installation
> method, like kickstart in Red Hat, and, for several reasons not worth
> discussing here, cannot build and burn custom CDs. I only have one PC, and
> want to fully customize it for my needs in the shortest possible time
> right after installation, from the default debian cd set.
dselect apt-get install <all the packages you need>
i would recommend dselect, because you get the recommendations for additional packages, too, not only the absolutely needed depends. But that could be against that, what you really want to do.
Daniel Kleine-Albers
> Which one of these methods is best for script use right after > installation, i.e.: > > No net access yet > Only the official debian CDs available > Minimum manual intervention (hopefully just > changing the CD when asked?)
First you will need to tell apt to use the CDROMS to find packages, apt-cdrom will take care of this for you.
Then just run, it will ask you to insert CDs when required.
apt-get install <package list>
> I am quite expert with red hat, and have started to study debian recently. > One of the things I found most confusing is the relationship betweeen apt, > dselect (?dpkg?) and what not. When do you use one or the other? Why?
For instance you wish to install vim, using apt-get install vim you will not get the vim-rt package which contains some other various goodies for use with vim. This is because the vim package doesn't require vim-rt to run, it just recommends it. apt-get install will only install the required packages.
dselect on the other hand is more user oriented. When you select vim, it will show you what is recommended and then get you to decide which of the packages you wish to install (ie vim-rt).
Both apt-get and dselect only download the packages onto your system. dpkg is used by both apt-get and dselect to install the downloaded packages.
Pete
> Which one of these methods is best for script use right after > installation, i.e.: > > No net access yet > Only the official debian CDs available > Minimum manual intervention (hopefully just > changing the CD when asked?)
Then you should use apt-get. If you install a bare debian system apt/dpkg asks you, how many questions it should ask during installation. If you set the priority to critical, you don't get almost any questions during installation.
If you want to script, i recommend using apt-get because thats the simpliest way. Simply make a script:
apt-get install <the packages you need>
there should be a parameter —assume-yes which causes even less questions to be asked. see man apt-move
> I am quite expert with red hat, and have started to study debian recently. > One of the things I found most confusing is the relationship betweeen apt, > dselect (?dpkg?) and what not. When do you use one or the other? Why? Are > there any online documents explaining this specific issues?
in short:
is the backend as rpm is on suse or redhat.
and dselect are frontends. With dselect you normally use apt-get as a middle-part between dselect itself and dpkg.
the frontends have in common that the "acquire" the packages automatically from cdrom, ftp, http, …
with dpkg you can only install packages you have as files directly with dpkg -i <packagefile.deb>
For further info have a look at:
Daniel Kleine-Albers
Newsgroups: gmane.linux.debian.user Date: Tue, 18 Oct 2005 12:28:00 -0400
On Sat, 15 Oct 2005 16:07:51 -0500, Kent West wrote:
> > " HOW TO REINSTALL DEBIAN FROM CD" > > Again, Debianites don't reinstall their systems (except in rare cases), > so it's not really a well-documented procedure. The short version is to > back up your /home directory (or leave that partition alone if it's a > separate partition), and then just reinstall like you're doing a new > install (making sure not to partition/format your /home partition if it > exists, or restoring it if it's not a separate partition). > > If you're having some sort of problem, let us know what it is and we can > probably help you fix it without a reinstall.
Debian was designed by the pros and for the pros. It is still true now. I.e., if fixing the problem cost 10 times than reinstall, the pros would think "Ah, that's challenging. I'm sure I can handle it"; whereas norms like me would say, "why bother, let me reinstall" — the fresh system will be up and running in less than an hour, stable and without any problem.
Unfortunately, there is no mechanism in Debian to let you reinstall easily. So you have to go through the "10 times" approach — it took me several days (>3) to fully downgrade from Debian testing to Debian stable, after a single update command… Oh, wait, I still have "17 not upgraded" packages…
T
Newsgroups: gmane.linux.debian.user Date: Tue, 18 Oct 2005 14:18:37 +0200
at home I run a server with actually to many stuff on it to be safe. But i want to test a lot of things that it evolved to this situation. It has ldap, samba, courier, spamassassin, clamav, squirrel, exim4,… Anyway, for normal quick restores, restoring a backup is quick and painless.
But i was thinking of a way to not only "restore" the system but also "move" it. For instance, my system acts weird sometimes because of my tampering with it so i would love to start from a clean sarge install but with all my data and services running on it without spending to much time on the reinstall.
To manage things better i've begun moving services to uml instances on the server and it keeps thing organised. Now this also makes restoring the services easy as one would only have to install a base system, install needed utils (uml,bridge) and copy the uml files containing the systems and start those.
As for other files like /etc/profile, /etc/inputrc, /etc/environment, i guess you could make a package containing those files and installing them when you install the "customizing" package or whatever you would call such a package.
A server reinstall or recovery would look like this:
install debian
install "mybase" package which would be a fake pacakge depending on real apps that you want to
install like aptitude less vim
install "base config files" package package containing config files, … /etc/profile /etc/inputrc /etc/skel /etc/environment /etc/bash.bashrc …
This would leave you with a server "customized" to your liking and with the base apps you can't do without. Then further customizing would be required to run the services:
install "uml base" package installs uml-utilities bridge-utils ` config files ` root file system /etc/init.d/uml-bridge /etc/init.d/uml-start /usr/local/uml/rootfs_template /usr/local/uml/rootfs
Next would be restoring the data
restore /root /home
As for backups, you would need to backup /root /home and the uml systems + maintain changes you make to files used in the custom packages.
As said, this is not the fastest restore method but with regards to reinstalling a server it might be pretty quick and versatile.
Is this doable? Any things i'm overlooking/comments/…
What would be an easy way to making such custom packages be it for installing config files or fake packages used to install your favourite apps?
Benedict Verheyen
> As said, this is not the fastest restore method but with regards to > reinstalling a server it might be pretty quick and versatile. > > 1. Is this doable? Any things i'm overlooking/comments/... > 2. What would be an easy way to making such custom packages be it for > installing config files or fake packages used to install your favourite > apps?
There are folks who have been thinking about this problem for a lot longer than you or me. Check out http://infrastructures.org for the concepts and ISConf[1] for the software.
Ryan Nowakowski
> There are folks who have been thinking about this problem for a lot
very interesting site. Wish I had know it earlier.
Just that there is not much said on the web. Can you give more links? Because skimming through it, I get a feeling that we are talking about apple and banners here. The OP's goal is a flexible restore/install system, but I think the site talks about a monster organization wide distributed cloning system.
,----- | It's not for use in environments where you want to still make manual or | ad-hoc changes to machines at the same time. | | * No gold server. You work from the command line of any representative | target machine. | | * No central repository. Packages and change orders are stored in a | distributed cache, checksummed, replicated, and spread across all | participating machines. | | * No CVS server. See the previous point. | | * No single point of failure. See above. | | * Better workflow. No more futzing around with CVS checkins, rsync updates, | or ssh'ing back to the gold server -- there isn't one. You log into one of | the machines you want to change -- more of a natural sysadmin workflow. `-----
T
> | * No gold server. You work from the command line of any representative > | target machine.
bad thing to have …. if it fails and there is no silver or bronze server with identical contents
> | * No central repository. Packages and change orders are stored in a > | distributed cache, checksummed, replicated, and spread across all > | participating machines.
repository is nice … if people need to be disciplined to comment all their changes before releasing to the rest of the machines
many hundred ways to do that "task" so that only tested files and patches gets out to the rest of the machines
> | * No CVS server. See the previous point.
ditto
> | * No single point of failure. See above.
ditto
> | * Better workflow.
if it takes more than a few minutes per day .. something is wrong
if the admin is afraid of 50 or 100 or 500 or 1000 or 5000 servers, something is wrong with the admin
the hardest part is to maintain and test the first 10-20 systems
and clone those patches/fixes onto the next 10-20 machines and the rest of the world picks up its changes from the "release" servers
> No more futzing around with CVS checkins, rsync updates,
that'd be a good or bad thing … depending on where you're looking
for small world of machines … say 5-25 …
make a cdrom with minimum drivers to support a network install off the net or clone any of your own local servers
it'd be more (geeky) fun to boot from (network) floppy and do a network install or even a usb-stick if you need more than 1.44/2.88MB to boot and install
* * if you lose your cdrom or floppy or usb-stick, * you better have a few backup or a way to recreate a new one *
or do a network boot for all machines, so they are all identical, except that the "pxe servers" will have to have a kernel that supports all the various client boxes
you cannot get away from "single point of failure"
fun stuff and ez to do …. in 5 min or 5hr or 5 days … would depend on what else *one* needs to do for the other 100 hrs of the week and what happens when any of the machine decides to go on holiday while you're on your honeymoon or vacation
alvin