New chapter: filesystems

Short overview

2 files changed, 261 insertions, 0 deletions

diff --git a/linux-basics/filesystems/Makefile b/linux-basics/filesystems/Makefile
new file mode 100644
index 0000000..4663d52
--- /dev/null
+++ b/linux-basics/filesystems/Makefile
@@ -0,0 +1,9 @@
+all:
+	for pdf in `ls -1 *.tex` ; do \
+		pdflatex $$pdf; \
+		pdflatex $$pdf; \
+	done
+
+clean:
+	rm -f *.aux *.log *.pdf *.log *.snm *.toc *.vrb *.nav *.out
+
diff --git a/linux-basics/filesystems/pres_filesystems_en.tex b/linux-basics/filesystems/pres_filesystems_en.tex
new file mode 100644
index 0000000..7dd5f57
--- /dev/null
+++ b/linux-basics/filesystems/pres_filesystems_en.tex
@@ -0,0 +1,252 @@
+\documentclass[11pt]{beamer}
+
+%\usepackage{ngerman}
+\usepackage{times}
+\usepackage{graphicx}
+\usepackage{pgf,pgfarrows,pgfnodes,pgfautomata,pgfheaps}
+\usepackage{amsmath,amssymb}
+\usepackage[latin1]{inputenc}
+\usepackage{listings,color}
+\definecolor{lbcolor}{RGB}{255,210,150}
+\lstset{
+	language=C,
+	numbers=left,
+	stepnumber=1,
+	numbersep=5pt,
+	numberstyle=\tiny,
+	breaklines=true,
+	breakautoindent=true,
+	postbreak=\space,
+	tabsize=2,
+	basicstyle=\ttfamily\footnotesize,
+	showspaces=false,
+	showstringspaces=false,
+	extendedchars=true,
+	backgroundcolor=\color{lbcolor},
+	keywordstyle=\bf ,
+	commentstyle=\color{green},
+	stringstyle=\color{red}
+}
+
+\mode<presentation>
+{
+	\usetheme{linutronix}
+}
+
+% on the following slides, include icon in the left sidebar
+\def\lximg{/usr/share/lx/icons/fueller.png}
+
+\title{Filesystems}
+\institute{Linutronix GmbH}
+
+\begin{document}
+
+\frame{ \titlepage }
+
+% stop displaying 'fueller.png' on the following slides
+\def\lximg{none}
+
+\begin{frame}
+\frametitle{Filesystem types}
+\begin{itemize}
+\item Filesystems for block devices
+\item Filesystems for flash devices
+\item Network filesystems (e.g. NFS)
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{Filesystems for block devices}
+\begin{itemize}
+\item EXT2
+\item EXT3, EXT4
+\item ReiserFS
+\item XFS
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{Flash filesystems}
+\begin{itemize}
+\item JFFS2
+\item YAFFS2
+\item LogFS
+\item UbiFS
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{Journalling Filesystems}
+\begin{alertblock}{What is a journalling filesystem?}
+A ''Journalling Filesystem'' keeps track of all changes in a special memory area
+(the journal). Based on the informations in the journal, files can be
+reconstructed if a write access gets interrupted (for example by a system
+crash or a power failure).
+\end{alertblock}
+\end{frame}
+
+\begin{frame}
+\frametitle{Technical terms}
+\begin{itemize}
+\item Block: Filesystems are usually organised in blocks. Small blocks won't
+waste that much memory for small files; large blocks will increase the maximum
+size of a filesystem
+\item Inode: Stores information about regular files, directories, ...
+\item Superblock: The superblock stores all information about the filesystem
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{EXT2}
+\begin{itemize}
+\item \textbf{N}o journalling filesystem
+\item A good choice for ''flash-based block devices'', like SD cards or USB
+sticks (ext2 minimizes the number of writes)
+\item Supported block sizes: 1kb, 2kb, 4kb, 8kb
+\item Max. filesystem size: 32TB (for 8kb blocks)
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{EXT3}
+\begin{itemize}
+\item \textbf{Journalling} filesystem
+\item Three levels of journalling available: journal, ordered, writeback
+\item Backward compatible with EXT2
+\item Limited to 32.000 sub-directories
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{EXT4}
+\begin{itemize}
+\item \textbf{Journalling} filesystem
+\item Backward compatible with EXT2 and EXT3
+\item Faster filesystem checking
+\item Filesystem size up 1 exabyte
+\item Journal checksumming
+\item \textbf{No longer} limited to 32.000 sub-directories
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{XFS}
+\begin{itemize}
+\item \textbf{Journalling} filesystem
+\item max. FILESIZE: 8 Exabyte !
+\item Online defragmentation (using xfs\_fsr)
+\item Online resizing (using xfs\_growfs)
+\item Good choice for servers (e.g. file servers)
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{JFFS2}
+\begin{itemize}
+\item Flash filesystem
+\item Used on top of MTD devices
+\item All nodes must be scanned at mount time
+\item Mount time and memory consumption linearly depend on the filesystem size
+\item \textbf{No} powerfail safety
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{UbiFS}
+\begin{itemize}
+\item Flash filesystem
+\item Used on top of UBI volumes
+\item Fast mount time
+\item Supports LZO and ZLIB compression
+\item Powerfail safety !!!
+\end{itemize}
+\end{frame}
+
+\begin{frame}
+\frametitle{Recommendations}
+\begin{itemize}
+\item ''Flash-based block devices'': EXT2
+\item Block devices: EXT3, EXT4
+\item Block devices which are used in server machines (on which huge amounts
+of data are stored): XFS
+\item Flash devices: UbiFS
+\end{itemize}
+\end{frame}
+
+\begin{frame}[fragile]
+\frametitle{Creating filesystems}
+\begin{lstlisting}[keywordstyle=\ttfamily\footnotesize]
+# EXT2 filesystem in a file
+$ dd if=/dev/zero of=ext2.img \
+     bs=2048 count=1024
+$ /sbin/mkfs.ext2 -b 2048 ext2.img 
+myext2.img is not a block special device.
+Proceed anyway? (y,n) y
+Filesystem label=
+OS type: Linux
+Block size=2048 (log=1)
+Fragment size=2048 (log=1)
+256 inodes, 1024 blocks
+51 blocks (4.98%) reserved for the super user
+First data block=0
+1 block group
+16384 blocks per group, 16384 fragments per group
+256 inodes per group
+\end{lstlisting}
+\end{frame}
+
+\begin{frame}[fragile]
+\frametitle{Creating filesystems}
+\begin{lstlisting}[keywordstyle=\ttfamily\footnotesize]
+# EXT2 filesystem on a disk partition
+mkfs.ext2 -b 2048 /dev/sdb2
+# Checking an EXT2 filesystem
+fsck.ext2 /dev/sdb2
+# Mounting an EXT2 partition
+mount -t ext2 /dev/sdb2 /mountpoint
+# or just
+mount /dev/sdb2 /mountpoint
+\end{lstlisting}
+\end{frame}
+
+\begin{frame}[fragile]
+\frametitle{Testing filesystems for block devices}
+\begin{lstlisting}[keywordstyle=\ttfamily\footnotesize]
+# 1) Mounting a filesystem which is located in a file
+mount -o loop ext2.imt /mountpoint
+# 2) Mounting a specific partition of a complete disk image
+$ fdisk -lu bla.img 
+255 heads, 63 sectors/track, 0 cylinders, total 0 sectors
+Units = sectors of 1 * 512 = 512 bytes
+Sector size (logical/physical): 512 bytes / 512 bytes
+[...]
+  Device Boot      Start         End      Blocks   Id  System
+bla.img1              63       64259       32098+  83  Linux
+# Now multiply the start offset (63) with
+# the sector size (512) = 32256
+$ mount -o offset=32256 /mountpoint
+\end{lstlisting}
+\end{frame}
+
+\begin{frame}
+\frametitle{Testing flash filesystems}
+\begin{itemize}
+\item You can emulate flash devices on your host system, if you want to test
+your flash filesystem
+\item \textbf{mtdram} can emulate NOR devices
+\item \textbf{nandsim} can emulate NAND devices
+\item This will be covered in a different chapter
+\end{itemize}
+\end{frame}
+
+\section*{sources}
+\begin{frame}
+\begin{thebibliography}{1}
+\bibitem{ext2} http://en.wikipedia.org/wiki/Ext2
+\bibitem{ext3} http://en.wikipedia.org/wiki/Ext3
+\bibitem{ext4} http://en.wikipedia.org/wiki/Ext4
+\bibitem{xfs} http://en.wikipedia.org/wiki/XFS
+\bibitem{UBIFS} http://en.wikipedia.org/wiki/UBIFS
+\end{thebibliography} 
+\end{frame}
+\end{document}