Difference between revisions of "Filebench"

Revision as of 19:38, 31 May 2011

Introduction

Filebench is a file system and storage benchmark that allows to generate a large variety of workloads. Unlike typical benchmarks it is very flexible and allows to minutely specify (any) applications' behaviour using extensive Workload Model Language (WML). Filebench uses loadable workload personalities to allow easy emulation of complex applications (mail servers, web servers, database server, etc.). Filebench is quick to set up and use compared to a real bulky applications. It is also a handy tool for micro-benchmarking.

Features

Filebench includes many features to facilitate file system benchmarking:

• Multiple workload types support via loadable personalities
• Includes pre-defined mail, web, file and database server personalities (as well as many others)
• Easy to add new personalities using Workload Model Language (WML)
• Reach WML language for workload specification (even includes synchronisation primitives for Database-like workloads)
• Multi-threaded and multi-process workload support
• Integrated statistics for throughput, latency, and CPU cycle counts per system call
• Fractal directory hierarchies with depth, width, and file sizes set to a given statistical distributions
• Tested on Linux, FreeBSD, and Solaris platforms and should work for any POSIX-compliant Operating System
• XXX: think anbout most relevant features

News

Filebench continues to evolve.

XXX: update with what hapened recently and what is in the plans. Releases.

Support

• If you have problems with using Filebench, use our Help Forum
• If you want to discuss some more generic questions about file system and storage benchmarking, use our Open Discussion Forum
• If you think that you found a BUG, please report it to our Bug Tracking System
• XXX: mention FAQ

Installing

Getting and Installing Filebench

Filebench is distributed as a compressed tar-ball with sources. Download the newest version of Filebench [here]. Uncompress it and follow the regular procedure for building UNIX sources:

• ./configure
• make
• sudo make install

If all of these steps complete successfully, you're ready to run Filebench. If you see any errors or warnings, please report them to Bug Tracking System.

Filebench does not have any mandatory program/library dependencies except libc. If you want command line auto-completion to work in Filebench you need to install libtecla prior to Filebench.

What is in the Package

The package installs into /usr/benchmarks/filebench and contains the following 6 subdirectories:

• bin: The binaries directory contains architecture specific subdirectories (amd64 etc) and binary stubs to execute the correct go_filebench binary via isaexec(5). filebench is a perl executeable and is the batch harness for go_filebench.
• docs: The douments directory contains READMEs, licensing information and this manual.
• config: This directory contains sample profiles used as input to filebench for running benchmarks and processing statistics in batches.
• scripts: The scripts directory contains utilities for gathering various different system statistics.
• workloads: Pre-defined f scripts representing micro and application level benchmarks of interest are contained in this directory.
• xanadu: This directory contains Xanadu which is a separate included tool bundle - a collection of utilities for gathering, processing and graphically displaying system statistics.

Running

Filebench applies file system or disk based load by executing "workload personalities", which define the workload to apply to the system, and provide various tunables to customize the workload. Filebench is shipped with a library of these personalities, ready to use. They are located in /usr/benchmarks/filebench/workloads.

There are two ways to run Filebench: by using the filebench wrapper script or by invoking the go_filebench core program. The filebench script is the easiest way to run workloads, especially if a suite of workload runs is desired. go_filebench can be used in order to do interactive running (this is typically useful for development testing).

Running with filebench

In the getting started section, the suggested command line:

# /usr/benchmarks/filebench/bin/filebench randomread

actually ran the filebench wrapper script which used a profile file in /usr/benchmarks/filebench/config named randomread.prof to specify the actual workload model to run and set some run specific defaults for it. The randomread.prof profile supplied with Filebench specifies the randomread.f workload and sets the directory for its test file to /tmp, the I/O size to two kilobytes, and total test file size to 160 megabytes. There are currently five profiles supplied with Filebench:

• randomread.prof - 2 KB random read using the randomread.f workload
• seqread.prof - Four streams of one megabyte sequential reads using the multistreamread.f workload.
• fileio.prof - A suite of benchmarks that measure basic I/O throughput
• filemicro.prof - A suite of "micro" benchmarks
• filemacro.prof - A suite of large and application oriented benchmarks.

All these profiles are located in the /usr/benchmarks/filebench/config directory.

Running interactively with go_filebench

Those that still need to access the executable (formerly called filebench) can do so by invoking go_filebench:

# /usr/benchmarks/filebench/bin/go_filebench

This will run Filebench in interactive mode, allowing loading and running of individual workload personalities, with full control over their parameters. The following example demonstrates interactively running the varmail workload personality.

Example varmail run

There are simple workload personalities, which configure the type of workload to simulate. An example is a /var/mail directory simulation (like postmark):

$ /usr/benchmarks/filebench/bin/go_filebench filebench> load varmail 8395: 3.898: Varmail personality successfully loaded 8395: 3.899: Usage: set $dir=<dir> 8395: 3.900: set $filesize=<size> defaults to 16384 8395: 3.900: set $nfiles=<value> defaults to 1000 8395: 3.901: set $dirwidth=<value> defaults to 20 8395: 3.901: set $nthreads=<value> defaults to 1 8395: 3.902: set $meaniosize=<value> defaults to 16384 8395: 3.902: run <runtime> filebench> set $dir=/tmp filebench> run 10 8395: 14.886: Fileset mailset: 1000 files, avg dir = 20, avg depth = 2.305865, mbytes=15 8395: 15.301: Preallocated fileset mailset in 1 seconds 8395: 15.301: Starting 1 filereader instances 8396: 16.313: Starting 1 filereaderthread threads 8395: 19.323: Running for 10 seconds... 8395: 29.333: Stats period = 10s 8395: 29.347: IO Summary: 21272 iops 2126.0 iops/s, (1063/1063 r/w) 32.1mb/s, 338us cpu/op, 0.3ms latency 8395: 29.348: Shutting down processes filebench> stats dump "stats.varmail" filebench> quit

This run did 21272 logical operations per second - e.g. 21272 open, close, read/write etc...

$ more stats.varmail Flowop totals: closefile4 492ops/s 0.0mb/s 0.0ms/op 31us/op-cpu readfile4 492ops/s 7.1mb/s 0.6ms/op 171us/op-cpu openfile4 492ops/s 0.0mb/s 5.4ms/op 653us/op-cpu closefile3 492ops/s 0.0mb/s 1.1ms/op 150us/op-cpu appendfilerand3 492ops/s 7.7mb/s 0.0ms/op 55us/op-cpu readfile3 492ops/s 7.2mb/s 0.6ms/op 163us/op-cpu openfile3 492ops/s 0.0mb/s 5.1ms/op 623us/op-cpu closefile2 492ops/s 0.0mb/s 0.9ms/op 136us/op-cpu appendfilerand2 492ops/s 7.7mb/s 0.1ms/op 67us/op-cpu createfile2 492ops/s 0.0mb/s 8.9ms/op 1151us/op-cpu deletefile1 492ops/s 0.0mb/s 8.4ms/op 1007us/op-cpu IO Summary: 649116 iops 5409.0 iops/s, 983/983 r/w 29.7mb/s, 1461uscpu/op

Workloads

Workload personalities define the workload to apply, and the default tunables for scaling that workload to the system.

Workloads are stored in workload definition files which Filebench loads, and can be easily customized from either the Filebench command line interface or by editing the files themselves (or copies of the files).

Filebench is shipped with a library of different workloads, which are described in the following sections.

Application Emulation

Application-Emulation Workloads Currently Available in Filebench package

varmail

A /var/mail NFS mail server emaulation, following the workload of postmark, but multi-threaded. The workload consists of a multi-threaded set of open/read/close, open/append/close and deletes in a single directory.

fileserver

A file system workload, similar to SPECsfs. This workload performs a sequence of creates, deletes, appends, reads, writes and attribute operations on the file system. A configurable hierarchical directory structure is used for the file set.

oltp

A database emulator. This workload performance transactions into a filesystem using an I/O model from Oracle 9i. This workload tests for the performance of small random reads & writes, and is sensitive to the latency of moderate (128k+) synchronous writes as in the log file. It by default launches 200 reader processes, 10 processes for asynchronous writing, and a log writer. The emulation includes use of ISM shared memory as per Oracle, Sybase etc which is critical to I/O efficiency (as_lock optimizations).

dss

DSS Database: this is future work TBD.

webserver

A mix of open/read/close of multiple files in a directory tree, plus a file append (to simulate the web log). 100 threads are used by default. 16k is appended to the weblog for every 10 reads.

webproxy

A mix of create/write/close, open/read/close, delete of multiple files in a directory tree, plus a file append (to simulate the proxy log). 100 threads are used by default. 16k is appended to the log for every 10 read/writes.

Micro-benchmark

copyfiles

A copy of a large directory tree. This workload creates a hierarchical directory tree, then measures the rate at which files can be copied from the source tree to a new tree. A single thread is used by default, although this is configurable.

createfiles

Create a directory tree and fill files. A populate of files of specified sizes a directory tree. File sizes are chosen according to a gamma distribution of 1.5, with a mean size of 16k.

randomread

A multi-threaded read of a single large file, defaulting to 8k reads. A single thread is used by default, although configurable by $nthreads.

randomwrite

A multi-threaded write of a single large file, defaulting to 8k writes. A single thread is used by default, although configurable by $nthreads.

singlestreamread

A sequential read of a large file. 1MB reads are used by default.

singlestreamwrite

A sequential write of a large file. 1MB writes are used by default.

multistreamwrite

A sequential read of 4 large files, each with their own reader thread. 1MB reads are used by default.

multistreamwrite

A sequential write of 4 large files, each with their own writer thread. 1MB writes are used by default.

Recommended parameters

The following recommmended parameters have been provided as suggestions; customize as appropriate.

Small Configurations

These may be suitable for servers with one or two disks.

Worklads	Suggested Parameters
fileserver bringover createfiles deletefiles varmail webproxy webserver	set $nfiles=50000
randomread singlestreamread multistreamread singlestreamwrite multistreamwrite	set $filesize=1g
oltp	set $filesize=1g

Large Configurations

These may be suitable for servers with 20+ disks.

Worklads	Suggested Parameters
fileserver bringover createfiles deletefiles mongo varmail webproxy webserver	set $nfiles=100000
randomread randomwrite	set $filesize=5g set $nthreads=256
filemicro_createrand filemicro_rread filemicro_rwritefsync singlestreamread multistreamread singlestreamwrite multistreamwrite	set $filesize=5g
oltp	set $filesize=5g
filemicro_create filemicro_delete filemicro_seqwrite filemicro_seqwriterand filemicro_writefsync	set $count= 100000

Filebench Batch Run Harness

The filebench program calls go_filebench to automatically run multiple configurations. Each config section corresponds to one workload file (.f)

The command is as follows:

# /bin/benchmarks/filebench/bin/filebench <profile name>

The statistics results are saved in the directory specified in the DEFAULT section.

Workload Model Language

The Workload Model Language defines the workload personalities. It is a concise text based language that has the capability to simulate a variety of complex workload patterns. Filebench is shipped with numerous examples in the workloads directory. These workload files have a .f extension.

If you don't find a appropriate workload model among those supplied with Filebench, you can write your own. See the Writing Workload Models page for a tutorial on how to write Filebench workload models, and the Workload Model Language page for a description of the language syntax. As Filebench is a work in progress, this is not yet a stable interface.

The following are examples of what the language currently looks like.

singlestreamwritedirect.f

The following is the most basic example of a workload file. This workload creates a file and writes to it from a single process and (by default) a single thread.

#
# Copyright 2005 Sun Microsystems, Inc. All rights reserved.
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License.
# See the file LICENSING in this distribution for details.
#
 
set $dir=/tmp
set $nthreads=1
set $iosize=1m
 
define file name=largefile1,path=$dir
 
define process name=seqwrite,instances=1
{
  thread name=seqwrite,memsize=10m,instances=$nthreads
  {
    flowop write name=seqwrite,filename=largefile1,iosize=$iosize,directio
    flowop bwlimit name=limit
  }
}
 
echo "Single Stream Write Version 1.7 2005/06/21 21:18:53 personality successfully loaded"
usage "Usage: set \$dir=<dir>"
usage "       set \$filesize=<size> defaults to $filesize"
usage "       set \$nthreads=<value> defaults to $nthreads"
usage "       set \$iosize=<value> defaults to $iosize"
usage " "
usage "       run runtime (e.g. run 60)"

randomread.f

This is the definition file for a random read workload.

Filebench randomread Workload

varmail.f

The following workload file simulates /var/mail activity on a mail server. Note the variety of file operations that can be performed, such as readwholefile and appendfilerand.

#
# Copyright 2005 Sun Microsystems, Inc. All rights reserved
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License.
# See the file LICENSING in this distribution for details
#
 
set $dir=/tmp
set $nfiles=1000
set $meandirwidth=1000000
set $filesize=16k
set $nthreads=16
set $meaniosize=16k
 
define fileset name=bigfileset,path=$dir,size=$filesize,entries=$nfiles,dirwidth=$meandirwidth,prealloc=80
 
define process name=filereader,instances=1
{
  thread name=filereaderthread,memsize=10m,instances=$nthreads
  {
    flowop deletefile name=deletefile1,filesetname=bigfileset
    flowop createfile name=createfile2,filesetname=bigfileset,fd=1
    flowop appendfilerand name=appendfilerand2,iosize=$meaniosize,fd=1
    flowop fsync name=fsyncfile2,fd=1
    flowop closefile name=closefile2,fd=1
    flowop openfile name=openfile3,filesetname=bigfileset,fd=1
    flowop readwholefile name=readfile3,fd=1
    flowop appendfilerand name=appendfilerand3,iosize=$meaniosize,fd=1
    flowop fsync name=fsyncfile3,fd=1
    flowop closefile name=closefile3,fd=1
    flowop openfile name=openfile4,filesetname=bigfileset,fd=1
    flowop readwholefile name=readfile4,fd=1
    flowop closefile name=closefile4,fd=1
  }
}
 
echo "Varmail Version 1.24 2005/06/22 08:08:30 personality successfully loaded"
usage "Usage: set \$dir=<dir>"
usage "       set \$filesize=<size> defaults to $filesize"
usage "       set \$nfiles=<value> defaults to $nfiles"
usage "       set \$nthreads=<value> defaults to $nthreads"
usage "       set \$meaniosize=<value> defaults to $meaniosize"
usage "       set \$meandirwidth=<size> defaults to $meandirwidth"
usage "(sets mean dir width and dir depth is calculated as log (width, nfiles)"
usage " dirdepth therefore defaults to dir depth of 1 as in postmark"
usage " set $meandir lower to increase depth beyond 1 if desired)"
usage " "
usage "       run runtime (e.g. run 60)"

List of Flowops

Here you can find the List of Flowops.

List of Attributes

Here you can find the List of Attributes.

Filebench for Programmers

See Filebench for Programmers for details on the Filebench code.

If you want to participate in Filebench development in any way, join Filebench Developers Mailing List