en/handbook/hb-install-mips-disk.xml

<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE sections SYSTEM "/dtd/book.dtd">

<!-- The content of this document is licensed under the CC-BY-SA license -->
<!-- See http://creativecommons.org/licenses/by-sa/2.5 -->

<!-- $Header: /var/cvsroot/gentoo/xml/htdocs/doc/en/handbook/draft/hb-install-mips-disk.xml,v 1.10 2006/02/26 15:17:34 fox2mike Exp $ -->

<sections>

<version>1.13</version>
<date>2006-02-27</date>

<section>
<title>Introduction to Block Devices</title>
<subsection>
<title>Block Devices</title>
<body>

<p>
We'll take a good look at disk-oriented aspects of Gentoo Linux
and Linux in general, including Linux filesystems, partitions and block devices.
Then, once you're familiar with the ins and outs of disks and filesystems,
you'll be guided through the process of setting up partitions and filesystems
for your Gentoo Linux installation.
</p>

<p>
To begin, we'll introduce <e>block devices</e>. The most famous block device is
probably the one that represents the first SCSI HD in a Linux system, namely
<path>/dev/sda</path>.
</p>

<p>
The block devices above represent an abstract interface to the disk. User
programs can use these block devices to interact with your disk without worrying
about whether your drives are IDE, SCSI or something else. The program can
simply address the storage on the disk as a bunch of contiguous,
randomly-accessible 512-byte blocks.
</p>

</body>
</subsection>
<subsection>
<title>Partitions</title>
<body>

<p>
Although it is theoretically possible to use a full disk to house your Linux
system, this is almost never done in practice. Instead, full disk block devices
are split up in smaller, more manageable block devices. These are called 
<e>partitions</e>. 
</p>

</body>
</subsection>
</section>
<section>
<title>Designing a Partitioning Scheme</title>
<subsection>
<title>How Many and How Big?</title>
<body>

<p>
The number of partitions is highly dependent on your environment. For instance,
if you have lots of users, you will most likely want to have your
<path>/home</path> separate as it increases security and makes backups easier.
If you are installing Gentoo to perform as a mailserver, your 
<path>/var</path> should be separate as all mails are stored inside 
<path>/var</path>. A good choice of filesystem will then maximise your 
performance. Gameservers will have a separate <path>/opt</path> as most gaming 
servers are installed there. The reason is similar for <path>/home</path>: 
security and backups. You will definitely want to keep <path>/usr</path> big:
not only will it contain the majority of applications, the Portage tree alone
takes around 500 Mbyte excluding the various sources that are stored in it.
</p>

<p>
As you can see, it very much depends on what you want to achieve. Separate
partitions or volumes have the following advantages:
</p>

<ul>
<li>
  You can choose the best performing filesystem for each partition or volume
</li>
<li>
  Your entire system cannot run out of free space if one defunct tool is
  continuously writing files to a partition or volume
</li>
<li>
  If necessary, file system checks are reduced in time, as multiple checks can
  be done in parallel (although this advantage is more with multiple disks than
  it is with multiple partitions)
</li>
<li>
  Security can be enhanced by mounting some partitions or volumes read-only, 
  nosuid (setuid bits are ignored), noexec (executable bits are ignored) etc.
</li>
</ul>

<p>
However, multiple partitions have one big disadvantage: if not configured 
properly, you might result in having a system with lots
of free space on one partition and none on another. There is also a 15-partition
limit for SCSI and SATA.
</p>

</body>
</subsection>
</section>
<section>
<title>Using fdisk on MIPS to Partition your Disk</title>
<subsection>
<title>SGI Machines: Creating an SGI Disk Label</title>
<body>

<p>
All disks in an SGI System require an <e>SGI Disk Label</e>, which serves a 
similar function as Sun &amp; MS-DOS disklabels -- It stores information about 
the disk partitions. Creating a new SGI Disk Label will create two special 
partitions on the disk:
</p>

<ul>
  <li>
    <e>SGI Volume Header</e> (9th partition): This partition is important. It
    is where the bootloader will reside, and in some cases, it will also contain
    the kernel images.
  </li>
  <li>
    <e>SGI Volume</e> (11th partition): This partition is similar in purpose to
    the Sun Disklabel's third partition of "Whole Disk". This partition spans
    the entire disk, and should be left untouched. It serves no special purpose
    other than to assist the PROM in some undocumented fashion (or it is used by
    IRIX in some way).
  </li>
</ul>

<warn>
The SGI Volume Header <e>must</e> begin at cylinder 0. Failure to do so means
you won't be able to boot from the disk.
</warn>

<p>
The following is an example excerpt from an <c>fdisk</c> session. Read and
tailor it to your needs...
</p>

<pre caption="Creating an SGI Disklabel">
# <i>fdisk /dev/sda</i>

Command (m for help): <i>x</i>

Expert command (m for help): <i>m</i>
Command action
   b   move beginning of data in a partition
   c   change number of cylinders
   d   print the raw data in the partition table
   e   list extended partitions
   f   fix partition order
   g   create an IRIX (SGI) partition table
   h   change number of heads
   m   print this menu
   p   print the partition table
   q   quit without saving changes
   r   return to main menu
   s   change number of sectors/track
   v   verify the partition table
   w   write table to disk and exit

Expert command (m for help): <i>g</i>
Building a new SGI disklabel. Changes will remain in memory only,
until you decide to write them. After that, of course, the previous
content will be irrecoverably lost.

Expert command (m for help): <i>r</i>

Command (m for help): <i>p</i>

Disk /dev/sda (SGI disk label): 64 heads, 32 sectors, 17482 cylinders
Units = cylinders of 2048 * 512 bytes

----- partitions -----
Pt#     Device  Info     Start       End   Sectors  Id  System
 9:  /dev/sda1               0         4     10240   0  SGI volhdr
11:  /dev/sda2               0     17481  35803136   6  SGI volume
----- Bootinfo -----
Bootfile: /unix
----- Directory Entries -----

Command (m for help):
</pre>

<note>
If your disk already has an existing SGI Disklabel, then fdisk will not allow
the creation of a new label. There are two ways around this. One is to create a
Sun or MS-DOS disklabel, write the changes to disk, and restart fdisk. The
second is to overwrite the partition table with null data via the following
command: <c>dd if=/dev/zero of=/dev/sda bs=512 count=1</c>.
</note>
</body>
</subsection>

<subsection>
<title>Getting the SGI Volume Header to just the right size</title>
<body>

<impo>
This step is often needed, due to a bug in <c>fdisk</c>.  For some reason, the
volume header isn't created correctly, the end result being it starts and ends
on cylinder 0.  This prevents multiple partitions from being created.  To get
around this issue... read on.
</impo>

<p>
Now that an SGI Disklabel is created, partitions may now be defined. In the
above example, there are already two partitions defined for you. These are the
special partitions mentioned above and should not normally be altered. However,
for installing Gentoo, we'll need to load a bootloader, and possibly multiple
kernel images (depending on system type) directly into the volume header.
The volume header itself can hold up to <e>eight</e> images of any size, 
with each image allowed eight-character names.
</p>

<note>
On SGI systems, there are two bootloaders available: <c>arcboot</c> and
<c>arcload</c>.  Currently, only the Indy, Indigo2 (R4k variety), Challenge S
and O2 are capable of using <c>arcboot</c> as a bootloader, whereas
<c>arcload</c> runs on all the Linux-supported SGI machines.  Unlike
<c>arcboot</c> however, <c>arcload</c> cannot presently read EXT2/3 partitions,
and thus, loads its kernels from the volume header.
</note>

<p>
The process of making the volume header larger isn't exactly straight-forward;
there's a bit of a trick to it. One cannot simply delete and re-add the volume
header due to odd fdisk behavior. In the example provided below, we'll create a
50MB Volume header in conjunction with a 50MB /boot partition. The actual layout
of your disk may vary, but this is for illustrative purposes only.
</p>

<pre caption="Resizing the SGI Volume Header correctly">
Command (m for help): <i>n</i>
Partition number (1-16): <i>1</i>
First cylinder (5-8682, default 5): <i>51</i>
 Last cylinder (51-8682, default 8682): <i>101</i>

<comment>(Notice how fdisk only allows Partition #1 to be re-created starting at a     )
(minimum of cylinder 5?  Had you attempted to delete &amp; re-create the SGI      )
(Volume Header this way, this is the same issue you would have encountered.   )
(In our example, we want /boot to be 50MB, so we start it at cylinder 51 (the )
(Volume Header needs to start at cylinder 0, remember?), and set its ending   )
(cylinder to 101, which will roughly be 50MB (+/- 1-5MB).                     )</comment>

Command (m for help): <i>d</i>
Partition number (1-16): <i>9</i>

<comment>(Delete Partition #9 (SGI Volume Header))</comment>

Command (m for help): <i>n</i>
Partition number (1-16): <i>9</i>
First cylinder (0-50, default 0): <i>0</i>
 Last cylinder (0-50, default 50): <i>50</i>

<comment>(Re-Create Partition #9, ending just before Partition #1)</comment>
</pre>

<p>
If you're unsure how to use <c>fdisk</c> have a look down further at the
instructions for partitioning on Cobalts.  The concepts are exactly the same --
just remember to leave the volume header and whole disk partitions alone.
</p>

<p>
Once this is done, you are safe to create the rest of your partitions as you see
fit. After all your partitions are laid out, make sure you set the partition ID
of your swap partition to <c>82</c>, which is Linux Swap. By default, it will be
<c>83</c>, Linux Native. 
</p>

<p>
Now that your partitions are created, you can now continue with <uri
link="#filesystems">Creating Filesystems</uri>.
</p>

</body>
</subsection>

<subsection>
<title>Cobalt Machines: Partitioning your drive</title>
<body>

<p>
On Cobalt machines, the BOOTROM expects to see a MS-DOS MBR, so partitioning the
drive is relatively straightforward -- in fact, it's done the same way as you'd
do for an Intel x86 machine.  <e>However</e> there are some things you need to
bear in mind.
</p>

<ul>
  <li>
    Cobalt firmware will expect <path>/dev/hda1</path> to be a Linux partition
    formatted <e>EXT2 Revision 0</e>.  <e>EXT2 Revision 1 partitions will NOT
    WORK!</e>  (The Cobalt BOOTROM only understands EXT2r0)
  </li>
  <li>
    The above said partition must contain a gzipped ELF image,
    <path>vmlinux.gz</path> in the root of that partition, which it loads as the
    kernel
  </li>
</ul>

<p>
For that reason, I recommend creating a ~20MB <path>/boot</path> partition
formatted EXT2r0 upon which you can install CoLo &amp; your kernels.  This
allows you to run a modern filesystem (EXT3 or ReiserFS) for your root
filesystem.
</p>

<p>
I will assume you have created <path>/dev/hda1</path> to mount later as a
<path>/boot</path> partition.  If you wish to make this <path>/</path>, you'll
need to keep the PROM's expectations in mind.
</p>

<p>
So, continuing on... To create the partitions you type <c>fdisk /dev/hda</c> at
the prompt.  The main commands you need to know are these:
</p>

<ul>
  <li>
    <c>o</c>: Wipe out old partition table, starting with an empty MS-DOS
    partition table
  </li>
  <li>
    <c>n</c>: New Partition
  </li>
  <li>
    <c>t</c>: Change Partition Type
    <ul>
      <li>Use type <c>82</c> for Linux Swap, <c>83</c> for Linux FS</li>
    </ul>
  </li>
  <li>
    <c>d</c>: Delete a partition
  </li>
  <li>
    <c>p</c>: Display (print) Partition Table
  </li>
  <li>
    <c>q</c>: Quit -- leaving old partition table as is.
  </li>
  <li>
    <c>w</c>: Quit -- writing partition table in the process.
  </li>
</ul>

<pre caption="Partitioning the disk">
# <i>fdisk /dev/hda</i>

The number of cylinders for this disk is set to 19870.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)

<comment>(Start by clearing out any existing partitions)</comment>
Command (m for help): <i>o</i>
Building a new DOS disklabel. Changes will remain in memory only,
until you decide to write them. After that, of course, the previous
content won't be recoverable.


The number of cylinders for this disk is set to 19870.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)
Warning: invalid flag 0x0000 of partition table 4 will be corrected by w(rite)

<comment>(You can now verify the partition table is empty using the 'p' command)</comment>

Command (m for help): <i>p</i>

Disk /dev/hda: 10.2 GB, 10254827520 bytes
16 heads, 63 sectors/track, 19870 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

   Device Boot      Start         End      Blocks   Id  System

<comment>(Create the /boot partition)</comment>

Command (m for help): <i>n</i>
Command action
   e   extended
   p   primary partition (1-4)
<i>p</i>
Partition number (1-4): <i>1</i>

<comment>(Just press ENTER here to accept the default)</comment>

First cylinder (1-19870, default 1):
Last cylinder or +size or +sizeM or +sizeK (1-19870, default 19870): <i>+20M</i>

<comment>(and now if we type 'p' again, we should see the new partition)</comment>
Command (m for help): <i>p</i>

Disk /dev/hda: 10.2 GB, 10254827520 bytes
16 heads, 63 sectors/track, 19870 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hda1               1          40       20128+  83  Linux

<comment>(The rest, I prefer to put in an extended partition, so I'll create that)</comment>

Command (m for help): <i>n</i>
Command action
   e   extended
   p   primary partition (1-4)
<i>e</i>
Partition number (1-4): <i>2</i>

<comment>(Again, the default is fine, just press ENTER.)</comment>

First cylinder (41-19870, default 41):
Using default value 41

<comment>(We want to use the whole disk here, so just press ENTER again)</comment>
Last cylinder or +size or +sizeM or +sizeK (41-19870, default 19870):
Using default value 19870

<comment>(Now, the / partition -- I use separate partitions for /usr, /var,
etc... so / can be small.  Adjust as per your preference.)</comment>

Command (m for help): <i>n</i>
Command action
   l   logical (5 or over)
   p   primary partition (1-4)
<i>l</i>
First cylinder (41-19870, default 41):<i>&lt;Press ENTER&gt;</i>
Using default value 41
Last cylinder or +size or +sizeM or +sizeK (41-19870, default 19870): <i>+500M</i>

<comment>(... and similar for any other partitions ...)</comment>

<comment>(Last but not least, the swap space.  I recommend at least 250MB swap,
preferrably 1GB)</comment>

Command (m for help): <i>n</i>
Command action
   l   logical (5 or over)
   p   primary partition (1-4)
<i>l</i>
First cylinder (17294-19870, default 17294): <i>&lt;Press ENTER&gt;</i>
Using default value 17294
Last cylinder or +size or +sizeM or +sizeK (1011-19870, default 19870): <i>&lt;Press ENTER&gt;</i>
Using default value 19870

<comment>(Now, if we check our partition table, everything should mostly be ship
shape except for one thing...)</comment>

Command (m for help): <i>p</i>

Disk /dev/hda: 10.2 GB, 10254827520 bytes
16 heads, 63 sectors/track, 19870 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

Device Boot      Start         End      Blocks      ID  System
/dev/hda1               1          21       10552+  83  Linux
/dev/hda2              22       19870    10003896    5  Extended
/dev/hda5              22        1037      512032+  83  Linux
/dev/hda6            1038        5101     2048224+  83  Linux
/dev/hda7            5102        9165     2048224+  83  Linux
/dev/hda8            9166       13229     2048224+  83  Linux
/dev/hda9           13230       17293     2048224+  83  Linux
/dev/hda10          17294       19870     1298776+  83  Linux

<comment>(Notice how #10, our swap partition is still type 83?)</comment>

Command (m for help): <i>t</i>
Partition number (1-10): <i>10</i>
Hex code (type L to list codes): <i>82</i>
Changed system type of partition 10 to 82 (Linux swap)

<comment>(That should fix it... just to verify...)</comment>

Command (m for help): <i>p</i>

Disk /dev/hda: 10.2 GB, 10254827520 bytes
16 heads, 63 sectors/track, 19870 cylinders
Units = cylinders of 1008 * 512 = 516096 bytes

Device Boot      Start         End      Blocks      ID  System
/dev/hda1               1          21       10552+  83  Linux
/dev/hda2              22       19870    10003896    5  Extended
/dev/hda5              22        1037      512032+  83  Linux
/dev/hda6            1038        5101     2048224+  83  Linux
/dev/hda7            5102        9165     2048224+  83  Linux
/dev/hda8            9166       13229     2048224+  83  Linux
/dev/hda9           13230       17293     2048224+  83  Linux
/dev/hda10          17294       19870     1298776+  82  Linux Swap

<comment>(Now, we write out the new partition table.)</comment>

Command (m for help): <i>w</i>
The partition table has been altered!

Calling ioctl() to re-read partition table.
Syncing disks.

#
</pre>

<p>
And that's all there is to it.  You should now be right to proceed onto the
next stage: <uri link="#filesystems">Creating Filesystems</uri>.
</p>

</body>
</subsection>
</section>

<section id="filesystems">
<title>Creating Filesystems</title>
<subsection>
<title>Introduction</title>
<body>

<p>
Now that your partitions are created, it is time to place a filesystem on them. 
If you don't care about what filesystem to choose and are happy with what we use
as default in this handbook, continue with <uri 
link="#filesystems-apply">Applying a Filesystem to a Partition</uri>.
Otherwise read on to learn about the available filesystems...
</p>

</body>
</subsection>
<subsection>
<title>Filesystems?</title>
<body>

<p>
Several filesystems are available.  ReiserFS, EXT2 and EXT3 are found stable
on the MIPS architectures, others are experimental.
</p>

<p>
<b>ext2</b> is the tried and true Linux filesystem but doesn't have metadata
journaling, which means that routine ext2 filesystem checks at startup time can
be quite time-consuming. There is now quite a selection of newer-generation
journaled filesystems that can be checked for consistency very quickly and are
thus generally preferred over their non-journaled counterparts. Journaled
filesystems prevent long delays when you boot your system and your filesystem
happens to be in an inconsistent state.
</p>

<p>
<b>ext3</b> is the journaled version of the ext2 filesystem, providing metadata
journaling for fast recovery in addition to other enhanced journaling modes like
full data and ordered data journaling. ext3 is a very good and reliable
filesystem. It has an additional hashed b-tree indexing option that enables 
high performance in almost all situations. You can enable this indexing by
adding <c>-O dir_index</c> to the <c>mke2fs</c> command. In short, ext3 is an 
excellent filesystem.
</p>

<p>
<b>ReiserFS</b> is a B*-tree based filesystem that has very good overall 
performance and greatly outperforms both ext2 and ext3 when dealing with small 
files (files less than 4k), often by a factor of 10x-15x. ReiserFS also scales 
extremely well and has metadata journaling. As of kernel 2.4.18+, ReiserFS is 
solid and usable as both general-purpose filesystem and for extreme cases such 
as the creation of large filesystems, the use of many small files, very large 
files and directories containing tens of thousands of files. 
</p>

<p>
<b>XFS</b> is a filesystem with metadata journaling which comes with a robust
feature-set and is optimized for scalability. We only recommend using this
filesystem on Linux systems with high-end SCSI and/or fibre channel storage and
an uninterruptible power supply. Because XFS aggressively caches in-transit data
in RAM, improperly designed programs (those that don't take proper precautions
when writing files to disk and there are quite a few of them) can lose a good
deal of data if the system goes down unexpectedly.
</p>

<p>
<b>JFS</b> is IBM's high-performance journaling filesystem. It has recently 
become production-ready and there hasn't been a sufficient track record to 
comment positively nor negatively on its general stability at this point.
</p>

</body>
</subsection>
<subsection id="filesystems-apply">
<title>Applying a Filesystem to a Partition</title>
<body>

<p>
To create a filesystem on a partition or volume, there are tools available for 
each possible filesystem:
</p>

<table>
<tr>
  <th>Filesystem</th>
  <th>Creation Command</th>
</tr>
<tr>
  <ti>ext2</ti>
  <ti><c>mke2fs</c></ti>
</tr>
<tr>
  <ti>ext3</ti>
  <ti><c>mke2fs -j</c></ti>
</tr>
<tr>
  <ti>reiserfs</ti>
  <ti><c>mkreiserfs</c></ti>
</tr>
<tr>
  <ti>xfs</ti>
  <ti><c>mkfs.xfs</c></ti>
</tr>
<tr>
  <ti>jfs</ti>
  <ti><c>mkfs.jfs</c></ti>
</tr>
</table>

<p>
For instance, to have the boot partition (<path>/dev/sda1</path> in our
example) in ext2 and the root partition (<path>/dev/sda3</path> in our example)
in ext3, you would use:
</p>

<pre caption="Applying a filesystem on a partition">
# <i>mke2fs /dev/sda1</i>
# <i>mke2fs -j /dev/sda3</i>
</pre>

<p>
Now create the filesystems on your newly created partitions (or logical
volumes).
</p>

<warn>
If you're installing on a Cobalt server, remember <path>/dev/hda1</path> MUST be
of type <e>EXT2 revision 0</e>; Anything else (e.g. EXT2 revision 1, EXT3,
ReiserFS, XFS, JFS and others) <e>WILL NOT WORK!</e>
You can format the partition using the command: <c>mke2fs -r 0 /dev/hda1</c>.
</warn>

<warn>
Also, be aware that arcboot currently is not able to read any filesystem other
than EXT2, EXT3 and ISO9660 (recent versions).  For that reason,
<path>/boot</path> on SGI machines must also reside on an EXT2 or EXT3 partition.
</warn>

</body>
</subsection>
<subsection>
<title>Activating the Swap Partition</title>
<body>

<p>
<c>mkswap</c> is the command that is used to create and initialize swap partitions:
</p>

<pre caption="Creating a Swap signature">
# <i>mkswap /dev/sda2</i>
</pre>

<p>
To activate the swap partition, use <c>swapon</c>:
</p>

<pre caption="Activating the swap partition">
# <i>swapon /dev/sda2</i>
</pre>

<p>
Create and activate the swap with the commands mentioned above.
</p>

</body>
</subsection>
</section>
<section>
<title>Mounting</title>
<body>

<p>
Now that your partitions are initialized and are housing a filesystem, it is
time to mount those partitions. Use the <c>mount</c> command. Don't forget to
create the necessary mount directories for every partition you created. As an
example we mount the root and boot partition:
</p>

<pre caption="Mounting partitions">
# <i>mount /dev/sda3 /mnt/gentoo</i>
# <i>mkdir /mnt/gentoo/boot</i>
# <i>mount /dev/sda1 /mnt/gentoo/boot</i>
</pre>

<note>
If you want your <path>/tmp</path> to reside on a separate partition, be sure to
change its permissions after mounting: <c>chmod 1777 /mnt/gentoo/tmp</c>. This
also holds for <path>/var/tmp</path>.
</note>

<p>
We will also have to mount the proc filesystem (a virtual interface with the 
kernel) on <path>/proc</path>. But first we will need to place our files on the partitions.
</p>

<p>
Continue with <uri link="?part=1&amp;chap=5">Installing the Gentoo
Installation Files</uri>.
</p>

</body>
</section>
</sections>