Sponsored Content
Special Forums Hardware Filesystems, Disks and Memory data from blktrace: read speed V.S. write speed Post 302365199 by achenle on Monday 26th of October 2009 09:42:31 AM
Old 10-26-2009
The write operations are almost certainly getting cached somewhere, most likely in the kernel page cache or the cache on the disk(s) themselves depending on your OS and hardware setup.
 

10 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

Speed it up!

I wonder, are there any "tricks" to increase my server's access time in general? (4 Replies)
Discussion started by: pappous
4 Replies

2. UNIX for Dummies Questions & Answers

Speed of mv vs. cp

Hi, Is mv (move) command quicker than cp (copy command)? I have large files and I want to know if mv actually copy the data to a new file then deletes the old or whether it just alters information the file system without physically moving data - Unfortuanately I don't have large files to test... (2 Replies)
Discussion started by: GMMike
2 Replies

3. Filesystems, Disks and Memory

dmidecode, RAM speed = "Current Speed: Unknown"

Hello, I have a Supermicro server with a P4SCI mother board running Debian Sarge 3.1. This is the "dmidecode" output related to RAM info: RAM speed information is incomplete.. "Current Speed: Unknown", is there anyway/soft to get the speed of installed RAM modules? thanks!! Regards :)... (0 Replies)
Discussion started by: Santi
0 Replies

4. Filesystems, Disks and Memory

Write Speed into a big file (in Gb's)

If a file size increases in Linux/UNIX to say in GB's then will there be a decrease in write speed. I mean will it take more time to write to a large file then to a small one?? Please clarify? Thanks in advance (2 Replies)
Discussion started by: anilgurwara
2 Replies

5. Shell Programming and Scripting

How to speed up grep?

hi i'm greping the files with sepefic keyword, where the file is of too big. Assume there are 10 days log file each of more than 200mb. i've to grep all those files with a specific keywords. for example, 1. i'll grep for error message 2. after the i'll do one more grep for keyword... (4 Replies)
Discussion started by: vij_krr
4 Replies

6. Filesystems, Disks and Memory

Disk read and write speed.

Being a novice user to linux i m little unaware of how would i check disk read write speed. One of my mate is suggesting to create a file using dd command and check how much time it takes to create a 30 gb file . I think this has a little sense however i would also like to take your reviews... (5 Replies)
Discussion started by: pinga123
5 Replies

7. Hardware

Hardware Correction: How to change DVD write speed

I am now on Kernel 2.6.32-26 For me 16x CD write speed is okay. I have old hardware which was able to write DVDs at 1x, back in previous linux version. Now, I dont get speed of less than 4x. Tested on k3b, xfburn, and brasero. But all start at bottom 4x write speed. k3b forced back to... (0 Replies)
Discussion started by: makh
0 Replies

8. Shell Programming and Scripting

Speed Up Grep

Hi, I have to grep string from 20 - 30 files each carries 200 - 300 MB size and append to the file. How to speed the grepping time. cat catalina.out_2012_01_01 | grep "xxxxx" >> backup.txt PLZ, Suggest me, Regards, Nanthagopal A (5 Replies)
Discussion started by: nanthagopal
5 Replies

9. Red Hat

Read speed too bad on NAS share on VM

Hi, We have two servers in scenario (vmsoldot01 is Oracle VM with Linux and tldtppod15 is physical Linux server). One NAS share is mounted on both servers with similar permissions and access. But READ speed is too bad on virtual in comparison to physical server. While trying to diagnose this, I... (2 Replies)
Discussion started by: solaris_1977
2 Replies

10. UNIX for Advanced & Expert Users

How to read a fast written log file at Real time speed?

Hello All, I am building a real time parser for a log file in my application. The log file is continuously written at a very fast pace and gets rolled over every 10 minutes. I have measured the speed and observed that around 1000 lines are written to it every second, each line about 30-40... (7 Replies)
Discussion started by: cool.aquarian
7 Replies
iowatcher(1)						      General Commands Manual						      iowatcher(1)

NAME
iowatcher - Create visualizations from blktrace results SYNOPSIS
iowatcher OPTIONS... DESCRIPTION
iowatcher graphs the results of a blktrace run. It can graph the result of an existing blktrace, start a new blktrace, or start a new blk- trace and a benchmark run. It can then create an image or movie of the IO from a given trace. iowatcher can produce either SVG files or movies in mp4 format (with ffmpeg) or ogg format (with png2theora). OPTIONS
--help Print a brief usage summary. -d, --device <device> Controls which device you are tracing. You can only trace one device at a time for now. It is sent directly to blktrace, and only needed when you are making a new trace. -D, --blktrace-destination <destination> Destination for blktrace. -p, --prog <program> Program to run while blktrace is run. -K, --keep-movie-svgs Keep the SVG files generated for movie mode. -t, --trace <file|directory> Controls the name of the blktrace file. iowatcher uses a dump from blkparse, so -t tries to guess the name of the corresponding per CPU blktrace data files if the dump file doesn't already exist. If you want more than one trace in a given graph, you can specify -t more than once. If a directory is specified, iowatcher will use the name of the directory as the base name of the dump file and all trace files found inside the directory will be processed. -l, --label <label> Sets a label in the graph for a trace file. The labels are added in the same order the trace files are added. -m, --movie [spindle|rect] Create a movie. The file format depends on the extension used in the -o filename.* option. If you specify an .ogv or .ogg exten- sion, the result will be Ogg Theora video, if png2theora is available. If you use an .mp4 extension, the result will be an mp4 video if ffmpeg is available. You can use any other extension, but the end result will be an mp4. You can use --movie=spindle or --movie=rect, which changes the style of the IO mapping. -T, --title <title> Set a title to be placed at the top of the graph. -o, --output <file> Output filename (default: trace.svg). -r, --rolling <seconds> Control the duration for the rolling average. iowatcher tries to smooth out bumpy graphs by averaging the current second with sec- onds from the past. Larger numbers here give you flatter graphs. -h, --height <height> Set the height of each graph -w, --width <width> Set the width of each graph -c, --columns <columns> Numbers of columns in graph output -x, --xzoom <min:max> Limit processed time range to min:max -y, --yzoom <min:max> Limit processed sectors to min:max -a, --io-plot-action <action> Plot given action (one of Q,D,C) in IO graph -P, --per-process-io Distinguish between processes in IO graph -O, --only-graph <graph> Add a single graph to the output (see GRAPHS). By default all the graphs are included, but with -O you get only the graphs you ask for. -O may be used more than once. -N, --no-graph <type> Remove a single graph from the output (see GRAPHS). This may also be used more than once. GRAPHS
Choices for -O and -N are: io, tput, latency, queue_depth, iops, cpu-sys, cpu-io, cpu-irq, cpu-user, cpu-soft EXAMPLES
Generate graph from the existing trace.dump: iowatcher -t trace.dump -o trace.svg Skip the IO graph: iowatcher -t trace.dump -o trace.svg -N io Only graph tput and latency: iowatcher -t trace.dump -o trace.svg -O tput -O latency Generate a graph from two runs, and label them: iowatcher -t ext4.dump -t xfs.dump -l Ext4 -l XFS -o trace.svg Run a fio benchmark and store the trace in trace.dump, add a title to the top, use /dev/sda for blktrace: iowatcher -d /dev/sda -t trace.dump -T 'Fio Benchmark' -p 'fio some_job_file' Make a movie from an existing trace: iowatcher -t trace --movie -o trace.mp4 iowatcher(1)
All times are GMT -4. The time now is 04:16 AM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy