(2 ratings)
Which Linux distribution should I use?
The differences between the various Linux distributions ("distros") are minor: the installation program, choice of the bundled applications and tools, arrangement of a few things on the hard drive. Regardless of your choice of distro, most of Linux is still at the same, and standard hard drive locations are used for essential items. Whichever distribution you decide to install, you will end up with essentially the same Linux.
- They are both very popular (both an advantage for a newbie and a testimony to their quality).
- They are both general-purpose distributions.
- They both come with relatively easy setup programs.
- Both Mandrake and RedHat contributions to Linux are "open software" (this means that all the software written by the packaging corporations and included on the distribution CDs is licensed under the General Public License, GPL, so that it can be legally copied, given away, reused, etc.).
- Both Mandrake and Fedora can be obtained cheaply or free if you don't care for commercial support. This is a consequence of (4).
- Mandrake was once
originally based on RedHat, so both distributions are quite similar.
Software packages for RedHat often work on Mandrake (and vice versa)
without problems. However, Mandrake is a bit more automatized and makes
a somewhat nicer desktop than RedHat. At the same time, Mandrake
sometimes is not as rock stable as RedHat.
The most recent distributions we recommend (December 2003) is Mandrake 9.2, and "Fedora Core 1" (the predesessor of Fedora was "RedHat 9.0"). These are both excellent distributions. Be sure to specify the most recent version if ordering your software from a dealer--many dealers like to clear their inventory by sending you an older version (this applies not only to Linux). Generally, development under Linux is fast, and you don't want to waste your time with older distributions. The authors of this guide have no connection to Mandrake or RedHat (or any other Linux distributor) whatsoever.
We tried Debian and we liked it very much. It was probably as easy as RedHat, but Debian seems less common (hence, being newbies, we picked up Mandrake or RedHat). The great benefit of Debian is that it is 100% non-commercial (put together by volunteer hackers, the true Linux way) and it probably most strictly adheres to Linux standards (it probably sets the standards too). Another great benefit is that Debian crams on their numerous distribution CDs thousands of tools and applications--easily much more than any other distribution. All these tools/apps are nicely "packaged" (for ease of installation) and tested for compatibility. This makes Debian distro look monumental, safe, conservative, and always somewhat outdated. So yes, we would not have a problem recommending Debian as a great general-purpose Linux distribution. Debian calls itself "The Universal Operating System" for a good reason. At any time, Debian carries 3 versions. (1) The "stable" version (sometimes called "potato"), and we would not recommend it, unless you are really paranoid on stability and don't mind quite outdated packages. (2) The Debian "testing" version (sometimes called "woody") is probably as stable as the latest RedHat, and more stable than your current Mandrake. It is much more up do date than Debian "stable". Debian Woody is the version we like. (3) If you don't mind occasional trouble, you can also the the third branch called "unstable", which is likely quite up-to-date.
S.u.S.E Linux distribution (http://www.suse.com) is very popular in Europe. It surely looks German--a solid, general-purpose distro with an easy setup and an excellent reputation. Many users swear by SuSe. We couldn't find cheap Suse CDs though but it appears you can download it (I cannot find a link). Their product includes propriatory additions that will satisfy enterprise-level need to interface some popular propriatory applications (MS Exchange, Cross-over office, etc).
Slackware seems to be favorite among "hard-core hackers" who like customizing scripts. We would have trouble recommending Slackware for Linux newbies unless the newbie likes to feel cool. Our reviewer Bill Staehle says: "The real 'reason' for a newbie to avoid Slackware is that it is much more command line oriented, and lacks some of the 'cutsie slick and drool' tools that the other distributions have." We received feedback from Linux newbies who use Slackware and it works very well for them. It seems that Slackware is relatively simple and cool because of the lack of automation. Therefore, with a bit of effort, a computer-literate administrator can actually understand what is going in her operating system (this is not something I can always say about Mandrake, or MS Windows for that matter).
Knoppix Linux (http://www.knoppix.net/) is another distribution worth consideration. The main point of Knoppix is that it is a "live distro", i.e., it can be booted from a CD, without installation. This is excellent for trying Linux (if you like it, you can also install Knoppix on the hard drive). It is also makes a perfect disaster-recovery tool (distro on a CD is also safe because no malicious program can do anything to your executables, and non-invasive for the local storage as required for post-mortem analyses). Knoppix is also useful if you have to work under Linux on sombody elses computer: you insert Knoppix CD into the CDdrive, and perhaps exteral storage on the usb port (for personal storage), and you are all set to work in your own environment. When done, you take your chips home. You can mount the local computer resources if you have to. Interesting tool.
Gentoo seems to have some strong following. In Gentoo (hearsay, never used it), they have a cool installation/upgrade system which does anything from sources (a local compilation is required). Long compilations can be joy to watch but, well, they can take time. The resulting executables are taylored to your hardware so they are perhaps smaller or faster than those on a more-standard "already cooked" (binary) Linux distribution.
Corel was once working on their own Linux distribution apparently geared towards a nice and easy platform to run the Corel suite of office applications: WordPerfect wordprocessor, QuattroPro spreadsheet, Corel Presentations, Paradox database, CorelDraw artist package.... The Corel Linux was based on the Debian distribution. It looked initially very promissing, but it is unclear to me what Corel has done with it (was paid by Microsoft to drop it?). In brief, Corel Linux is dead now, and I would never recommend it to anybody because it it a dead-end. The only reason to mention it here is that Corel Linux once received lots of publicity, so you may still hear about it. It seems like a sad story, particularly for Canadians.
Caldera was once another well-known distribution. It was said to be aiming at corporate users, had a fancy (and pricey) configuration tools, and other corporate goodies. In Aug.2000, Caldera purchased SCO Unix (the original trademarked ancient UNIX) which gave them an even more "corporate" look in my eyes. Caldera did not seem to care too much about home Linux users, so I never considered it for my home use. In early 2003, Caldera (renamed SCO) evidently swiched to different, perhaps more profitable, business model ("fire programmers, hire lawyers"). I will surely stay clear of anything that might bear the name SCO or Caldera on it because I do not like the idea of paying US$1399 for the right to run Linux on a single-processor computer or being sued. Caldera/SCO Linux distro certainly does not have any future.
There are "localized" versions of Linux for specific countries or languages (Korean, Chinese, Japanese, Hungarian, French, ... )--they likely contain (on default) all the hacks and docs (documentation) that the users in these countries want to see. Says Bill Staehle: "You may want to mention the Conectiva Linux distribution, loosely based on RH from Brazil. As such, it is in Portugese, and is also available in Spanish. Try: http://www.conectiva.com.br. I heard several good things about Conectiva, so if Portugese or Spanish was my language, I would probably give it a try.
The distribution you need is of course specific to the hardware platform you have. This means that for your PC hardware containing an Intel 386 processor, or Intel 486, or Intel Pentium, or Intel 586, or Intel 686, or Cyrix, or K6, AMD, or similar, you need the binary distribution called "Intel" or "386" or x86. [Unless you are prepared to start with your own compilation of the Linux source code, which is not typical for a newbie :-)] . This happens because there are binary distributions for other hardware platforms too: PowerPC, Alpha, Apple, IBM mainframe, "Intel StrongARM", Transmeta, and perhaps a dozen more--you don't want to get those binaries for your PC clone; they surely will not work on a PC machine with an "Intel" or "AMD" processor inside. If you have no-Intel hardware, you may want to search the Internet to find who supports it (chances are Debian does, they seem to support even the most exotic ones. Then, you need to obtain "Debian ARM" or "Debian Motorola 680x0"or "Debian PowerPC" or "Debian SPARC ", ...).
In short, although newbies get confused with the multiple Linux distributions, there are reasons to have different distros. They should be viewed as a Linux strength rather than weakness. Linux is simply filling all application and hardware platform niches. The drawback is that there are some "funny" distribution to avoid if you plan your serious business to depend on Linux.
This guide concentrates on RedHat and Mandrake for the PC (Intel) platform. Many of the answers will work fine on other distributions or platforms, but we did not try them.
Which Linux Distribution should I select for my old computer(s)? Quick answer: Debian, Slackware, or perhaps BasicLinux (current version), or an older version of RedHat, Mandrake, or SuSE. Justification: RedHat, Mandrake, SuSE, Caldera, and TurboLinux are optimized/suitable for hardware current at the date of their release. They may be difficult or impossible to install on older machines mostly due to the memory contraints and speed. Debian and Slackware are suitable for most older hardware as well. At the begining of installation on older hardware, watch for the initial prompt to start the installation in the "text mode".
1.2 What are the Linux hardware requirements?
My 486-33 MHz with 8 MB memory and 1 GB hard drive had too little memory to run adequately stand-alone under GUI, but was still useful in my home network environment running as an X-terminal (a 486-class machine also performs just adequately stand-alone if it has at least 16 MB of memory but sometimes memory for old computers is hard to obtain and high-priced). (Pls note that Mandrake requires a Pentium processor.)
Networking is where Linux really shines, so consider getting 10-base-T Ethernet cards--they are not very expensive and will be perfect to connect your two or more home computers together. Also, look around for old Ethernet cards which MS Windows deems obsolete--they can be bought for a really low price and they will work great under Linux. The 10-base-T system uses "giant phone" (RJ45)-type connectors and normal (not cross-over) "Ethernet" cables. All the computers are connected, from the socket on the Ethernet card, to a small box called the hub. The hub has an extra connection (called "uplink") which I will use if I ever have a permanent "over-Ethernet" connection to the outside world. Here is a schematics for a straight-forward home network arrangement:
-----------
| The_Hub |-[uplink]---[to_External_Network_over_Ethernet]
-----------
| | |
PC1 PC2 PC3_with_modem---External_Network_over_PPP
Here, I show a local private network consisting of PC1, PC2 and PC3, connected through a hub. Since I do not have "External_Network_over_Ethernet" on my home hub "uplink", PC3 provides my connection to the outside world (over a modem). Therefore, PC3 is called the "gateway" for all computers on my local private network (except PC3 itself). I enable the firewall software on PC3 PPP network interface, and let PC3 know how to dial out and connect to the outside. The outside world can only see PC3. As far as they can tell, PC1 and PC2 do not exist. My local ethernet network is "trusted" because only trusted people have physical access to PC1, PC2 and PC3. ("PPP" stands for "Point-to-Point Protocol" and it is a standard for communicating over phone lines.)
You can, of course, build a more complicated network with Linux. A PC can have 2 (or more) ethernet cards. It may then work as bridge between 2 (or more) networks. The PC will act as a gataway for all traffic between between the 2 networks. The networks do not have to be known to the outside world ("local private networks") and sit behind a firewall enabled on a gateway computer. The outside world will only know about 1 computer of mine, the "gateway" to the external network. Other computers will still be able to communicate with the outside world, but all the traffic will appear outside to originate from one, very busy computer--the gateway.
To connect just two computers, a cross-over cable for direct Ethernet-card to Ethernet-card connection is sufficient ("networking for the poor") but the connection through the hub is more straight-forward. To connect more than 2 computers together, you need a hub (~US$30 to US$80) and normal (not cross-over) cables. (If you have extra Ethernet cards, you may also consider installing more than 1 Ethernet card on a computer to make it a router. Then you use direct connections to other computers using the cross-over cables, and save the expense of a hub. This adds a configuration complexity to your system, yet in some situations can be technically advantagous.)
Here is another suggestion on setting up a different kind of network, using a very much older type hardware, which uses coaxial cables (like for the cable TV). For this, no hub is necessary. Because this networking scheme is older, it can be assembled using cards and parts that are sometimes available for free:
(edited for space) From: John.Edwards@brunel.ac.uk Subject: Linux Guide-a suggestion
Hi. Many older 10Mbps network cards (and some newer ones as well) have a BNC connector and you can usually pick up old co-axial cabling when companies upgrade to UTP. Add a T piece for each machine and a 50-ohm terminator at each end (about 1 pound or $1.50 each) and you have a home network that will happily support more machines than you probably have room for. And most importantly--no expensive hub (or cheap hub that can cause trouble). There are other advantages to co-ax as well, it's tougher to break and more resistant to noise from other equipment.
Disadvantages: There is a limit of 185 meters per network segment of thin co-ax, 30 machines per network, and you're stuck at 10Mbps, but I don't see any small home network needing more than that. Also if one cable goes down then the whole network stops, this shouldn't happen often unless someone unplugs a cable section. You can disconnect the T piece from a PC without harming the rest though.
Quick diagram, T for a T piece and Term for a terminator:
Term-T-----T-----T-----T-Term
| | | |
PC PC PC PC
The various parts connect together using BNC connectors similar to a TV & video connector but with a bayonet that secures the two sockets together. For more details see the /usr/doc/HOWTO/Ethernet-HOWTOThe most straigh-forward and modern, however, is to get one 10-base-T ethernet card for each of your computers and a one hub to connect them all.
1.3 Will my hardware work under Linux?
The most notable exceptions are the so-called Winmodems (=MS Windows modems also called "software modems"). Avoid these like fire--they are a bit less expensive than full modems, but they are crippled (some processing is done by the main computer CPU instead of by the modem), and there is little chance you will have a Winmodem running on Linux right away (for more info on Winmodems, see http://www.idir.net/~gromitkc/winmodem.html). External modems are never "Winmodems" so if in doubt, purchase an external modem (external modems are more expensive, but they don't drain your PC power supply, are easily portable between machines, look better, and show modem activity). Additional points to consider with modems: "Older externals using a Rockwell Protocol that don't work too well. Also, the newer USB modems are not currently (March 2001) well supported. See the winmodem page." [source: B.Staehle].
Another area of potential problems is the video card. If you have a recent "cutting edge" 3D or uncommon card, you may want to check its compatibility at http://www.Xfree86.org.
Zip drives of all kinds are supported fine.
So the short answer is yes, in all likelihood your standard PC will run Linux with no problems. You don't invest much when trying Linux, so probably the easiest way to make sure is to attempt an installation on your existing hardware. There are Linux hardware compatibility lists at http://hardware.redhat.com/hcl/ and http://metalab.unc.edu/LDP/HOWTO/Hardware-HOWTO.html if you want to check your newer or less popular hardware.
If you are an adventurous person, as I am, I would pay no attention to the remarks above, chances are 90-10 that the hardware will work.
Another area that can be problematic is laptops, because laptops typically contain quite unusual hardware that requires the support of the original vendor. See http://www.linux.org/hardware/laptop.html for a help with installation. IBM laptops appear a safe bet as far as Linux compatiblity is concerned. To purchase a Linux laptop, you may also try http://www.linuxcertified.com/linux_laptops.html
1.4 How do I download Linux?
If you do have a speedy Internet connection (definitely not a 56k-modem, but perhaps cable modem) and you are not a complete newbie, a Linux download may be an option to you after all. Try http://www.linuxiso.org/ for ready-to-burn CD images (the ISO format) of your selected Linux distribution. To avoid problems, check the downloaded ISO image prior to burning using the MD5 sum using (on Linux):
md5sum the_name_of_the.iso
This command will take some time (a few minutes?) to complete a check of a typical (640 kByte) CD image. (MS Windows does not contain a utility to check the md5sum, but such a utility can be downloaded for free.)
The utility md5 generates a 32-character string which depends on the file contents. If even one bit in the file has changed, the string is going to be different. Once the above command generates that signature string (the "md5 sum"), compare this string with that listed on the web page from which you downloaded the file. Please note that data can be corrupted during large downloads, so checking the sum is really a good idea else you risk problems with installation.
1.5 How do I get a Linux CD?
Here is a more comprehensive list of places to obtain Linux CDs with their location, so you can find something near to you (after B.Staehle) : http://www.ixsoft.de (low price CDs in Europe); http://linuxservice.de (another source in Germany); http://www.polo.demon.co.uk/emporium.html (The Linux Emporium); http://www.linux-emporium.co.uk; http://definite.ukpost.com/ (Definite Linux Systems); http://www.mallind.demon.co.uk/ (GPL + official distros); http://www.amush.cx/linux/ (GPL distributions in UK only); http://www.linux123.co.uk/ (GPL + official distros in .uk); http://www.kihi.com.au/bowtie/ (Bowtie Software - cheap CDs in OZ); http://EverythingLinux.com.au/ (cheap CDs in OZ); http://www.lsl.com.au/ (cheap CDs in OZ); http://www.arles.ns.ca (official distros and BSD in .ca); http://www.affinity-systems.ab.ca/ (Official distros - hardware in .ca); http://www.warpedsystems.sk.ca (GPL + official distros, custom built system CA); http://www.softcopy.on.ca/ (cheap CDs in CA); http://www.libranet.com (Libranet Linux Vancouver, BC, CA); http://www.linuxwarehouse.co.za/ (low cost and official in South Africa); http://linuxcentral.com (Clinton Township MI 48035); http://www.lsl.com (Chesterfield, MI 48047, USA); http://www.cheapbytes.com (Lodi, CA 95241, USA); http://www.infomagic.com (Flagstaff AZ 86004, USA); http://www.tummy.com/krud/ (Fort Collins, CO 80525, USA); http://www.pieceby.com/ (Hudson, NH 03051, USA); http://www.linuxmall.com (Aurora, CO 80046-0190, USA); http://www.linux-now.com (Clarion, PA 16214, USA); http://www.xcomputing.com (San Francisco, CA 94134, USA); http://www.ccsoft.cc/linux/ (Santa Rosa, CA 95401, USA); http://www.storeanywhere.com/ (Brooklyn, NY 11235, USA); http://www.linuxcomponents.com (Owings Mills, MD 21117, USA).
1.6 I have Linux Installation CDs but no install floppy. What do I do?
If you don't know how to access your BIOS setup, read this paragraph. The BIOS setup can typically be entered at boottime by pressing the proper key at the right moment (often when a prompt is briefly displayed). Most often, it is the <Del> key. Here is a list of key combinations used by popular BIOSes: Acer notebooks: <F2> during Power-On Self-Test (POST). American Megatrends (AMI): <Del> during Power-On Self-Test. Award: <Del>, or <Ctrl><Alt><Esc>. Compaq: <F10> after the square appears in the top right corner of the screen during boot-up. Dell: <Ctrl><Alt><Enter>. DTK: <Esc> during Power-On Self-Test. IBM Aptiva 535: <F1> while the square with the wavy lines is displayed in the upper right corner during power-on. IBM PS/2: <Ctrl><Alt><Del>, then <Ctrl><Alt><Ins> when the cursor is in the top right corner Mr. BIOS: <Ctrl><Alt><S> during Power-On Self-Test. Packard Bell: For some models, <F1> or <F2 > during Power-On Self-Test. Phoenix: <Ctrl><Alt><Esc> or <Ctrl><Alt><S>, or <Ctrl><Alt><Enter>.
If your computer cannot boot from the CD drive, make an install boot diskette from under DOS or the MS Windows DOS mode. (You have go to to "Shutdown" and "Restart in MS-DOS mode", not just run a DOS window).
It is important that you have a perfectly good floppy (without even one bad cluster). The program that makes the diskette does not check if the floppy was written properly. Also, don't count on the DOS FORMAT utility finding a faulty floppy--it probably won't. If I were you, I would make two or three boot floppies at once--you may be surprised how many diskettes have problems. For me, the third floppy worked! If your install diskette does not boot, make another one--it definitely should boot.
Here are the commands. To make the boot floppy run:
F:\dosutils\rawrite.exe -f F:\images\boot.img -d a: -n
To make the supplemental (optional) diskette run:
F:\dosutils\rawrite.exe -f F:\images\supp.img -d a: -n
This assumes your CDROM is the DOS "F:" drive, and your floppy is "A:", adjust the commands if the drive letters are different on your system.
The
commands above run the utility "rawrite" and specify the input file
("disk image", after the option "-f") and the destination drive (after
the option "-d"), and suppresses the prompt to insert a floppy (option
"-n"). You may find it easier to run rawrite without any
argument--it will interactively prompt you for the input image (pick
the file name as in the commands above) and the destination drive
letter.
From under Linux, you can make a boot floppy as follows.
If the floppy needs low-level formatting, you can format it under Linux using:
fdformat /dev/fd0H1440
mount /mnt/cdrom
cd /mnt/cdrom/images/
Write the floppy image to the floppy:
dd if=boot.img of=/dev/fd0
[The ``dd'' command copies files. The above command specifies that the input file ("if") is ``boot.img'' and the output file ("of") is /dev/fd0, which is the first floppy drive, i.e. the floppy drive number zero (if you want to write to your second floppy drive, use /dev/fd1). ]
To verify that the disk image was written to the floppy correctly, you can use one of the following four ways:
[1st way]
cmp boot.img /dev/fd0
[2nd way]
diff boot.img /dev/fd0
[3rd way]
md5sum boot.img
md5sum /dev/fd0
[compare if the two returned md5 checksums are identical]
[4th way]
sum boot.img
sum /dev/fd0
[comare if the two returned sums are identical]
1.7 What do I need to read before installation?
You may want to learn about your hardware: how many and what size hard drives you have, the type, number, order and size of all partitions on each drive, where your DOS/Windows partitions are, which one is the DOS/Windows boot partition (if you plan to have dual boot), what type of mouse you have, what video card and with how much memory, what monitor (max synchronization frequencies), etc.
Go to BIOS setup to see the number and geometry of your hard drives. Run DOS "fdisk" to display your hard drive(s) partition table(s), and print it. Watch your system boot to learn about the type of your video card and the amount of video memory. Boot MS Windows, go to the control panel-devices and write down the sound card, modem, network card types and settings (name, type, IRQ, i/o address, DMA channel). Read the label underneath your mouse to see the type of mouse you have. (Next time you buy a mouse, get a Linux-ready 3-button Logitech or similar--Linux makes good use of all three buttons.) Dust off your monitor manual to find out the maximum synchronization frequencies (vertical and horizontal) that your monitor supports. Never use frequencies out of the monitor specification--this may damage your monitor.
You may want to browse the RedHat or Mandrake manual. If you don't have the printed copy, an html version is on your CD so you can read it using any web browser, e.g. Netscape for Windows. Look here to see how to access this manual and some additional reading material which is on your Linux CD.
1.8 Can I have MS Windows and Linux installed on the same computer?
From under Linux, you will be able to read from and write to your MS Windows drive partitions so that the data exchange between MS Windows- and the Linux-based program is seamless. You will also be able to use your existing MS Windows-based resources: sound files, backgrounds, pictures, fonts, etc. (First check if it does not violate your license agreement though, smile. For products that are on rent to you from Microsoft, it probably does. With my Linux computers, I can be proud to have no pirated software on my system whatsoever.)
MS Windows will have no notion that Linux exists on your computer--it will just see that
* a boot manager is present,
* a non-Microsoft partition(s) is present on the hard drive.
MS Windows does not interfere with neither of them during normal operations (but mendles badly during installation, hard-drive re-partitioning, etc.)
1.9 How do I partition my hard drive?
If you plan a dual boot (Linux and MS Windows on the same computer), first use your DOS/Win utility FDISK
to make the MS Windows partition(s). Leave part (half?) of the hard
drive(s) unpartitioned for Linux. You will create and format the Linux
partitions during your RedHat (or Mandrake or whatever else)
installation. Linux will recognize the free (unpartitioned) space on
the hard drive and use it to create partition(s) in a non-Microsoft
format.
Make the MS Windows partition "primary" and "bootable". Install, configure, and test your MS Windows before Linux installation. If you plan to run Linux only (no dual boot), you need just a clean hard drive (no partitions) to start with.
During the Linux setup, you will be asked to partition the available space on your hard drive(s). There are many possible ways to partition, depending on your hard drive space, requirements, and taste. I like Linux hard drive partitions like this (for a modest total of 2 GB of hardrive space which I give to Linux in this example):
mount point type sizeIn the above example, I dedicate 300 MB for the root partition that holds the base of the Linux operating system. I allocate 1200 MB to the mount point that will be visible on my filesystem as the /usr directory and will contain the user's programs (the programs that don't come with the base operating system and I install later, for example StarOffice). I dedicate 380 MB for the partition that will be visible as the directory /home and will contain the setting and data of all users on the machine. And I allocate 120 MB to a "raw" partition for the operating system to use as the virtual memory (extension of the physical, silicon memory on the hard drive, so-called swap). (If your Linux is ancient, kernel version lower than 2.2 as in RH5.2 and earlier, your swap partition cannot be larger than approximately 127 MB.) The rule of thumb is that the swap should be about twice the amount of the physical memory (RAM). If you need more (e.g. if you have lots of physical memory, or you expect to run custom programs with really large data structures) you might want to create a larger swap partition during the installation (or several smaller swap partitions) or add a swap file(s) later.
/ ext2 300 MB
/usr ext2 1200 MB
/home ext2 380 MB
swap swap 120 MB
2 GB is a respectable amount of disk space and should be sufficient for users who like having many applications. (This is because Linux applications tend to be slimmer than their MS Windows equivalents). However, if you try to install everything that's available on the modern distribution CDs, you will surely run out of disk space. My experience is that however large the hard drive space, it will get filled and I regret I don't have more :-) .
If my space on the hard drive is really restricted, I may consider a two-partition setup like this (for a lean 650 MB total dedicated to Linux):
mount point type sizeIn this example, I dedicate 600 MB to the base of the operating system, applications, and user documents/data, and allow 50 MB for the swap partition (for the operating system to use as the virtual memory). The 50-MB swap should be quite sufficient for medium duty operations. The limitation of 600 MB for the operating system, applications and user data means that you will have to be very selective as to which applications you install or else you risk running out of hard drive space. Try pressing <F1> when installing the optional software that comes on the Red Hat CD--it will give you a short description of what the software does so you could perhaps decide if you really need it. (Don't worry too much if you miss something you need, you can install the missing parts later). You can easily finish the RedHat installation with 200 MB free on your Linux partition (out of 600 MB used in this example) if you make reasonable choices. Please note that "bundling together" the root partition "/" and the /home directory will likely save you some disk space, but it is not the safest solution.
/ ext2 600 MB
swap swap 50 MB
It is possible to install Linux on even less disk space than in the example above, but you will have to be really picky as to what you install.
For a larger available hard drive space, I nowadays (year 2002) consider the following setup (for a comfortable total of 15 GB dedicated to Linux):
mount point type sizePlease note that the the mount points can reside on different physical hard drives. Linux agglomarates all the hard drive space into a single directory tree.
/ ext2 800 MB
/usr ext2 5000 MB
/usr/local ext2 3000 MB
/home ext2 5200 MB
swap swap 1000 MB
Another consideration when setting up the partitions on older computers (486?). Many older BIOSes have the restriction that the boot partition cannot extend beyond the 1024th cylinder on your first physical hard drive. To overcome this limitation, simply make the first (bootable) partition so that it ends before the cylinder number 1023 (this makes this partition max approximately 512 MB in size, which is plenty for the "/" root partition). Once Linux boots, the BIOS restriction does not matter any more because Linux takes over the hardware management and it can access the partition(s) beyond the cylinder number 1023.
When installing and using Linux, your drives appear as devices with the following names: hda--first IDE drive (stands for "hard drive a", i.e. the master drive on the first IDE interface), hdb--second IDE drive (i.e., the slave drive on the first IDE interface), hdc--third IDE drive (i.e. the master drive on the second IDE interface), hdd--fourth IDE drive (i.e. the slave drive on the second IDE interface). The numbers mean the partitions on the physical drives: "hda1" means the first IDE hard drive (hd a), first partition (1); "hda2" is the first IDE hard drive, second partition; "hda3"--the first IDE hard drive, third partition; (and so on if you have more than 3 partitions on the first IDE hard drive); "hdb1"--second IDE hard drive, first partition (or just "hdb" if it is the CDROM installed as a slave on your first IDE interface). "hdc1"--third IDE hard drive, first partition, etc. SCSI drives have analogous names but start with the letters "sd" (="SCSI drive"), followed by the letter indicating the SCSI interface and by the number indicating the SCSI device id. For example, "sda4" means "first SCSI interface, id number 4". If you have an external zip drive attached to your parallel port, it will appear as SCSI device "sda4" (zip drives work in a SCSI-emulation mode).
The
listing of partitions that your Linux setup program presents to you
during installation will include any MS Windows partitions which you
have. For example, I have the following MS Windows partition:
mount point type size comment
[no mount] vfat 1200 MB ["Win C drive, hda1]
/mnt/dos_hdd2 vfat 1600 MB ["Win D drive, hdd2]
Don't erase these MS Windows partitions during your Linux installation if you want a dual boot. If you erase the MS Windows partition, MS Windows is gone from your system! If not sure, backup your data from your MS Windows partitions before Linux installation. "msdos", "fat" and "vfat" and "ntfs" are typical filesystems used by DOS and MS Windows 3.x/95/98/NT.
minix is the filesystem used in the Minix operating system, the first to run under Linux. It has a number of shortcomings: a 64MB partition size limit, short filenames, a single time stamp, etc. It remains useful for floppies and RAM disks.
ext is an elaborate extension of the minix filesystem. It has been completely superseded by the second version of the extended filesystem (ext2) and will eventually be removed from the kernel.
ext2 is the high performance disk filesystem used by Linux for fixed disks as well as removable media. The second extended filesystem was designed as an extension of the extended file system (ext). ext2 offers the best performance (in terms of speed and CPU usage) of the filesystems supported under Linux. In short, ext2 is the main (default, typical) Linux filesystem.
ext3 is an extension of the ext2 filesystem with journaling. It is backwards and forward compatibile with ext2. It means that ext2 can be converted into ext3 without reformatting or data loss (just re-mounting the partion is required). ext3 can be changed back to ext2, also without data loss. I use ext3 extensively since Oct.2001--it is simple and trouble-free. It is included as an installation "option" since RedHat 7.2 and Mandrake 8.0. It is highly recommended that you use this filesystem.
xiafs was designed and implemented to be a stable, safe filesystem by extending the Minix filesystem code. It provides the basic most requested features without undue complexity. The xia filesystem is no longer actively developed or maintained. It is used infrequently.
msdos is the filesystem used by DOS, Windows, and some OS/2 computers. msdos filenames can be no longer than 8 characters followed by an optional period and 3 character extension.
umsdos is an extended DOS filesystem used by Linux. It adds capability for long filenames, UID/GID, POSIX permissions, and special files (devices, named pipes, etc.) under the DOS filesystem, without sacrificing compatibility with DOS.
vfat is an extended DOS filesystem used by Microsoft Windows95 and Windows NT. VFAT adds capability for long filenames under the MSDOS filesystem.
proc is a pseudo-filesystem which is used as an interface to kernel data structures rather than reading and interpreting /dev/kmem. In particular, its files do not take up disk space. See man 5 proc.
iso9660 is a CD-ROM filesystem type conforming to the ISO 9660 standard. Two extensions (listed below) are automatically supported.
High Sierra --Linux supports High Sierra, the precursor to the ISO 9660 standard for CD-ROM filesystems. It is automatically recognized within the iso9660 filesystem support under Linux.
Rock Ridge --Linux also supports the System Use Sharing Protocol records specified by the Rock Ridge Interchange Protocol. They are used to further describe the files in the iso9660 filesystem to a UNIX host, and provide information such as long filenames, UID/GID, POSIX permissions, and devices. It is automatically recognized within the iso9660 filesystem support under Linux.
hpfs is the High Performance Filesystem, used in OS/2. This filesystem is read-only under Linux due to the lack of available documentation.
sysv is an implementation of the SystemV/Coherent filesystem for Linux. It implements all of Xenix FS, SystemV/386 FS, and Coherent FS.
nfs is the network filesystem used to access disks located on remote computers.
smb is a network filesystem that supports the SMB protocol, used by MS Windows for Workgroups, Windows NT, and Lan Manager. To use smb fs, you need a special mount program, which can be found in the ksmbfs package, found at ftp://sunsite.unc.edu/pub/Linux/system/Filesystems/smbfs. [Standard linux command "smbmount" will also do.]ncpfs is a network filesystem that supports the NCP protocol, used by Novell NetWare.
devpts--is a pseudo file system, traditionally mounted on /dev/pts. In order to acquire a pseudo terminal, a process opens /dev/ptmx; the number of the pseudo terminal is then made available to the process and the pseudo terminal slave can be accessed as /dev/pts/<number>.
fat --is not a separate filesystem, but a common part of the msdos, umsdos and vfat filesystems.
UFS --is a file system widely used
20 Random Tutorials from the same category :
