Posts Tagged ‘Disk’

Monitoring Linux hardware Hard Drives / Temperature and Disk with lm_sensors / smartd / hddtemp and Zabbix Userparameter lm_sensors report script

Thursday, April 30th, 2020

monitoring-linux-hardware-with-software-temperature-disk-cpu-health-zabbix-userparameter-script

I'm part of a  SysAdmin Team that is partially doing some minor Zabbix imrovements on a custom corporate installed Zabbix in an ongoing project to substitute the previous HP OpenView monitoring for a bunch of Legacy Linux hosts.
As one of the necessery checks to have is regarding system Hardware, the task was to invent some simplistic way to monitor hardware with the Zabbix Monitoring tool.  Monitoring Bare Metal servers hardware of HP / Dell / Fujituse etc. servers  in Linux usually is done with a third party software provided by the Hardware vendor. But as this requires an additional services to run and sometimes is not desired. It was interesting to find out some alternative Linux native ways to do the System hardware monitoring.
Monitoring statistics from the system hardware components can be obtained directly from the server components with ipmi / ipmitool (for more info on it check my previous article Reset and Manage intelligent  Platform Management remote board article).
With ipmi
 hardware health info could be received straight from the ILO / IDRAC / HPMI of the server. However as often the Admin-Lan of the server is in a seperate DMZ secured network and available via only a certain set of routed IPs, ipmitool can't be used.

So what are the other options to use to implement Linux Server Hardware Monitoring?

The tools to use are perhaps many but I know of two which gives you most of the information you ever need to have a prelimitary hardware damage warning system before the crash, these are:
 

1. smartmontools (smartd)

Smartd is part of smartmontools package which contains two utility programs (smartctl and smartd) to control and monitor storage systems using the Self-Monitoring, Analysis and Reporting Technology system (SMART) built into most modern ATA/SATA, SCSI/SAS and NVMe disks

Disk monitoring is handled by a special service the package provides called smartd that does query the Hard Drives periodically aiming to find a warning signs of hardware failures.
The downside of smartd use is that it implies a little bit of extra load on Hard Drive read / writes and if misconfigured could reduce the the Hard disk life time.

 

linux:~#  /usr/sbin/smartctl -a /dev/sdb2
smartctl 6.6 2017-11-05 r4594 [x86_64-linux-4.19.0-5-amd64] (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Device Model:     KINGSTON SA400S37240G
Serial Number:    50026B768340AA31
LU WWN Device Id: 5 0026b7 68340aa31
Firmware Version: S1Z40102
User Capacity:    240,057,409,536 bytes [240 GB]
Sector Size:      512 bytes logical/physical
Rotation Rate:    Solid State Device
Device is:        Not in smartctl database [for details use: -P showall]
ATA Version is:   ACS-3 T13/2161-D revision 4
SATA Version is:  SATA 3.2, 6.0 Gb/s (current: 3.0 Gb/s)
Local Time is:    Thu Apr 30 14:05:01 2020 EEST
SMART support is: Available – device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status:  (0x00) Offline data collection activity
                                        was never started.
                                        Auto Offline Data Collection: Disabled.
Self-test execution status:      (   0) The previous self-test routine completed
                                        without error or no self-test has ever
                                        been run.
Total time to complete Offline
data collection:                (  120) seconds.
Offline data collection
capabilities:                    (0x11) SMART execute Offline immediate.
                                        No Auto Offline data collection support.
                                        Suspend Offline collection upon new
                                        command.
                                        No Offline surface scan supported.
                                        Self-test supported.
                                        No Conveyance Self-test supported.
                                        No Selective Self-test supported.
SMART capabilities:            (0x0002) Does not save SMART data before
                                        entering power-saving mode.
                                        Supports SMART auto save timer.
Error logging capability:        (0x01) Error logging supported.
                                        General Purpose Logging supported.
Short self-test routine
recommended polling time:        (   2) minutes.
Extended self-test routine
recommended polling time:        (  10) minutes.

SMART Attributes Data Structure revision number: 1
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x0032   100   100   000    Old_age   Always       –       100
  9 Power_On_Hours          0x0032   100   100   000    Old_age   Always       –       2820
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       –       21
148 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       0
149 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       0
167 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       0
168 Unknown_Attribute       0x0012   100   100   000    Old_age   Always       –       0
169 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       0
170 Unknown_Attribute       0x0000   100   100   010    Old_age   Offline      –       0
172 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       –       0
173 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       0
181 Program_Fail_Cnt_Total  0x0032   100   100   000    Old_age   Always       –       0
182 Erase_Fail_Count_Total  0x0000   100   100   000    Old_age   Offline      –       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       –       0
192 Power-Off_Retract_Count 0x0012   100   100   000    Old_age   Always       –       16
194 Temperature_Celsius     0x0022   034   052   000    Old_age   Always       –       34 (Min/Max 19/52)
196 Reallocated_Event_Count 0x0032   100   100   000    Old_age   Always       –       0
199 UDMA_CRC_Error_Count    0x0032   100   100   000    Old_age   Always       –       0
218 Unknown_Attribute       0x0032   100   100   000    Old_age   Always       –       0
231 Temperature_Celsius     0x0000   097   097   000    Old_age   Offline      –       97
233 Media_Wearout_Indicator 0x0032   100   100   000    Old_age   Always       –       2104
241 Total_LBAs_Written      0x0032   100   100   000    Old_age   Always       –       1857
242 Total_LBAs_Read         0x0032   100   100   000    Old_age   Always       –       1141
244 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       32
245 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       107
246 Unknown_Attribute       0x0000   100   100   000    Old_age   Offline      –       15940

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log structure revision number 1
No self-tests have been logged.  [To run self-tests, use: smartctl -t]

Selective Self-tests/Logging not supported

 

2. hddtemp

 

Usually if smartd is used it is useful to also use hddtemp which relies on smartd data.
 The hddtemp program monitors and reports the temperature of PATA, SATA
 or SCSI hard drives by reading Self-Monitoring Analysis and Reporting
 Technology (S.M.A.R.T.)
information on drives that support this feature.
 

linux:~# /usr/sbin/hddtemp /dev/sda1
/dev/sda1: Hitachi HDS721050CLA360: 31°C
linux:~# /usr/sbin/hddtemp /dev/sdc6
/dev/sdc6: KINGSTON SV300S37A120G: 25°C
linux:~# /usr/sbin/hddtemp /dev/sdb2
/dev/sdb2: KINGSTON SA400S37240G: 34°C
linux:~# /usr/sbin/hddtemp /dev/sdd1
/dev/sdd1: WD Elements 10B8: S.M.A.R.T. not available

 

 

3. lm-sensors / i2c-tools 

 Lm-sensors is a hardware health monitoring package for Linux. It allows you
 to access information from temperature, voltage, and fan speed sensors.
i2c-tools
was historically bundled in the same package as lm_sensors but has been seperated cause not all hardware monitoring chips are I2C devices, and not all I2C devices are hardware monitoring chips.

The most basic use of lm-sensors is with the sensors command

 

linux:~# sensors
i350bb-pci-0600
Adapter: PCI adapter
loc1:         +55.0 C  (high = +120.0 C, crit = +110.0 C)

 

coretemp-isa-0000
Adapter: ISA adapter
Physical id 0:  +28.0 C  (high = +78.0 C, crit = +88.0 C)
Core 0:         +26.0 C  (high = +78.0 C, crit = +88.0 C)
Core 1:         +28.0 C  (high = +78.0 C, crit = +88.0 C)
Core 2:         +28.0 C  (high = +78.0 C, crit = +88.0 C)
Core 3:         +28.0 C  (high = +78.0 C, crit = +88.0 C)

 


On CentOS Linux useful tool is also  lm_sensors-sensord.x86_64 – A Daemon that periodically logs sensor readings to syslog or a round-robin database, and warns of sensor alarms.

In Debian Linux there is also the psensors-server (an HTTP server providing JSON Web service which can be used by GTK+ Application to remotely monitor sensors) useful for developers
psesors-server

psensor-linux-graphical-tool-to-check-cpu-hard-disk-temperature-unix

If you have a Xserver installed on the Server accessed with Xclient or via VNC though quite rare,
You can use xsensors or Psensora GTK+ (Widget Toolkit for creating Graphical User Interface) application software.

With this 3 tools it is pretty easy to script one liners and use the Zabbix UserParameters functionality to send hardware report data to a Company's Zabbix Sserver, though Zabbix has already some templates to do so in my case, I couldn't import this templates cause I don't have Zabbix Super-Admin credentials, thus to work around that a sample work around is use script to monitor for higher and critical considered temperature.
Here is a tiny sample script I came up in 1 min time it can be used to used as 1 liner UserParameter and built upon something more complex.

SENSORS_HIGH=`sensors | awk '{ print $6 }'| grep '^+' | uniq`;
SENSORS_CRIT=`sensors | awk '{ print $9 }'| grep '^+' | uniq`; ;SENSORS_STAT=`sensors|grep -E 'Core\s' | awk '{ print $1" "$2" "$3 }' | grep "$SENSORS_HIGH|$SENSORS_CRIT"`;
if [ ! -z $SENSORS_STAT ]; then
echo 'Temperature HIGH';
else 
echo 'Sensors OK';
fi 

Of course there is much more sophisticated stuff to use for monitoring out there


Below script can be easily adapted and use on other Monitoring Platforms such as Nagios / Munin / Cacti / Icinga and there are plenty of paid solutions, but for anyone that wants to develop something from scratch just like me I hope this
article will be a good short introduction.
If you know some other Linux hardware monitoring tools, please share.

30 years anniversary of the first mass produced portable computer COMPAQ Grid Compass 1011

Thursday, July 19th, 2012

Grid Notebook Big screen logo

Today it is considered the modern laptop (portable computers) are turning 30 years old. The notebook grandparent is a COMPAQGRiD Compass 1011 – a “mobile computer” with a electroluminescent display (ELD) screen supporting resolution of 320×240 pixels. The screen allowed the user to use the computer console in a text resolution of 80×24 chars. This portable high-tech gadget was equipped with magnesium alloy case, an Inten 8086 CPU (XT processor) at 8Mhz (like my old desktop pravetz pc 😉 ), 340 kilobyte (internal non-removable magnetic bubble memory and even a 1,200 bit/s modem!

COMPAQ Grid Compass considered first laptop / notebook on earthy 30 anniversary of the portable computer

The machine was uniquely compatible for its time as one could easily attach devices such as floppy 5.25 inch drives and external (10 Meg) hard disk via IEEE-488 I/O compatible protocol called GPiB (General Purpose instrumental Bus).

First mass prdocued portable computer laptop grid COMPAQ 11011 back side input peripherals

The laptop had also unique small weight of only 5 kg and a rechargable batteries with a power unit (like modern laptops) connectable to a normal (110/220 V) room plug.

First notebook in World ever the COMPAQ grid Compass 1101,br />
The machine was bundled with an own specificly written OS GRiD-OS. GRID-OS could only run a specialized software so this made the application available a bit limited.
Shortly after market introduction because of the incompitablity of GRID-OS, grid was shipped with MS-DOS v. 2.0.
This primitive laptop computer was developed for serve mainly the needs of business users and military purposes (NASA, U.S. military) etc.

GRID was even used on Space Shuttles during 1980 – 1990s.
The price of the machine in April 1982 when GriD Compass was introduced was the shockingly high – $8150 dollars.

The machine hardware design is quite elegant as you can see on below pic:

 COMPAQ grid laptop 1101 bubbles internal memory

As a computer history geek, I’ve researched further on GRID Compass and found a nice 1:30 hour video telling in detailed presentation retelling the history.

Shortly after COMPAQ’s Grid Compass 1011’s introduction, many other companies started producing similar sized computers; one example for this was the Epson HX-20 notebook. 30 years later, probably around 70% of citizens on the globe owns a laptop or some kind of portable computer device (smartphone, tablet, ultra-book etc.).

Most of computer users owning a desktop nowdays, owns a laptop too for mobility reasons. Interestengly even 30 years later the laptop as we know it is still in a shape (form) very similar to its original predecessor. Today the notebook sales are starting to be overshadowed by tablets and ultra-books (for second quarter laptop sales raised 5% but if compared with 2011, the sales rise is lesser 1.8% – according to data provided by Digital Research agency). There are estimations done by (Forrester Research) pointing until the end of year 2015, sales of notebook substitute portable devices will exceed the overall sales of notebooks. It is manifested today the market dynamics are changing in favour of tabets and the so called next generation laptopsULTRA-BOOKS. It is a mass hype and a marketing lie that Ultra-Books are somehow different from laptops. The difference between a classical laptop and Ultra-Books is the thinner size, less weight and often longer battery use time. Actually Ultra-Books are copying the design concept of Mac MacBook Air trying to resell under a lound name.
Even if in future Ipads, Android tablets, Ultra-Books or whatever kind of mambo-jambo portable devices flood the market, laptops will still be heavily used in future by programmers, office workers, company employees and any person who is in need to do a lot of regular text editting, email use and work with corporative apps. Hence we will see a COMPAC Grid Compass 1011 notebook likes to be dominant until end of the decade.

Viewing JPEG,GIF and PNG in ASCII with cacaview on GNU / Linux – Review on caca-utils text mode graphics utilities

Thursday, May 17th, 2012

Stitch 80x45 libcaca mascot cacaview viewing JPG, PNG, GIF images as ASCII on Linux libcaca

Probably, many don't know that it is possible to view normal graphical pictures (JPG, PNG, GIF, BMP) etc. in plain console tty.

Being able to view pictures in ASCII is something really nice especially for console geeks like me.
The images produced sometimes are a bit unreadable, if compared to the original graphics, but anyways most of the pictures looks pretty decent 🙂

Viewing in console / terminal images on GNU / Linux is possible thanks to a library called libcaca, caca labs libcaca project official website here.
Below is a shot description of libcaca:
hipo@noah:~$ apt-cache show libcaca0|grep 'Description' -A 4
Description: colour ASCII art library
libcaca is the Colour AsCii Art library. It provides high level functions
for colour text drawing, simple primitives for line, polygon and ellipse
drawing, as well as powerful image to text conversion routines.

In Debian, Ubuntu and other deb Linux distros viewing GUI images with no need for Xserver or any kind of window manager in plain ASCII is possible with cacaview.

cacaview is part of a package called caca-utils. caca-utils is providing few other great utilities for ASCII freaks 🙂 along with cacaview console ascii viewer prog.
The package> is available for Debian distributins since many years, so even on a very old Debians like Debian – (Potato, Woody, Sarge) the package is available in default free package repositories ready to install via apt

To install apt-get it as usual:

noah:~# apt-get --yes install caca-utils

Here is a list of the binaries the package provides:

hipo@noah:~$ dpkg -L caca-utils|grep -i /usr/bin/
/usr/bin/cacaserver
/usr/bin/cacaplay
/usr/bin/cacafire
/usr/bin/cacademo
/usr/bin/cacaview
/usr/bin/img2txt

1. cacaserver a tiny program allowing network streaming of applications written in caca

Belkow is a chop, from man cacaserver
 

cacaserver reads libcaca animation files in its standard input and
serves them as ANSI art on network port 51914. These animations can be
created by any libcaca program by setting the CACA_DRIVER environment
variable to raw and piping the program's standard output to cacaserver.

Clients can then connect to port 51914 using telnet or netcat to see
the output.

The example section of the manual points 1 example use of cacaserver to stream the console output from cacademo.
cacademo binary is a short presentation ASCII DEMO in the spirit of the old school assembly demos (demoscene) .
To run it to bind on port 51914 one has to type in bash shell:
hipo@noah:~$ CACA_DRIVER=raw cacademo | cacaserver
initialised network, listening on port 51914

Then to check out how the demo looks, open telnet connection to the cacaserver host; In my case the cacaserver is binded and streamed over IP 192.168.0.2:

hipo@debian:~$ telnet 192.168.0.2 51914

Immediately you got the demo shining; Below are two screenshots of the demo played after succesful telnet connection:

Cacaserver - caca for the network screenshot Matrix cacademo

cacademo running over telnet network connection – Matrix

cacaserver running on Debian GNU / Linux drug addict like spots streamed via telnet

Blur spots cacademo shot of cacademo streamed via network

You see the demo looks quite awesome 🙂

2. Running cacafire to stream over network

Another possible example use of cacaserver is in conjunction with cacafire libcaca test application:

noah:~# CACA_DRIVER=raw cacafire | cacaserver
initialised network, listening on port 51914
cacafire is a short application written to render ASCII via libcaca and is just displaying a screen with ASCII (moving) burning fire.
It is quite spectacular if you, ask an unexpecting friend to connect to your host to 51914 🙂

Cacafire Screenshot Debian GNU / Linux cacaserver streaming ASCII demo via network port 51914

Besides that bored sys admins, could run cacafire in console to hypnotize themselves watching dumb the burning fire screen for few hoursor just use it as a screensaver 😉

3. cacaview a program to display a graphic images in console using ASCII art

cacaview takes just one argument – the picture to be displayed.

Below is a screenshot of cacaview ran from my gnome-terminal displaying a ASCII text version of the MySQL server logo

hipo@noah:~$ cd /disk/pictures
hipo@noah:/disk/pictures$ cacaview mysql_logo.png

 

cacaview displaying MySQL database logo in ASCII using caca for X

Whether cacaview is invoked in GUI, the libcaca X support is used, so the text image is visualized in new window with graphics, if however it is invoked in plain let's say tty1 libcaca displays the graphics pictures drawing it with only text characters.

Here is also a screenshot, I've made while viewing a GIF website logo in ASCII in plain tty console:

hipo@noah:~$ cacaview /disk/pictures/logo.gif

cacaview plain tty console screenshot of a website logo graphics pictures 17-05-2012

The logo is in cyrillic, so for latin speaking people some of the characters in the two words seen will be unreadable 🙂

cacaview even supports viewing, the next and previous picture in line, like in any modern graphics image viewer program.
To view a bunch of graphic pictures in ASCII with cacaview pass it *.*:

hipo@noah:~$ cacaview /disk/pictures/*.*

For simplicity the common unix * is also supported, so I find it quicker to do:

hipo@noah:~$ cacaview /disk/pictures/*

Showing pictures forward and backward (Previous / Next) picture is done with n and p kbd keys, whether;
n - next;
p - previous

cacaview doesn't crash or stop but skip unknown file formats – if for instance encounters filenames which are not images; lets say you have *.rar archive files along with other pictures.

The complete list of keys cacaview supports are:
br />

KEYS
? show the help screen

n, p switch to next image, previous image

Left, Right, Up, Down or h, l, k, j
scroll the image around

+, – zoom in and out

z reset the zoom level to normal

f switch fullscreen mode (hide/show menu and status bars)

d toggle the dithering mode (no dithering, 4×4 ordered dithering, 8×8 ordered dithering and random dithering)

q exit the program

4. Converting graphics images to ASCII art like (plain text pictures)

The tool that does "the trick" is img2txt. img2txt has a bit more options while compared to the rest of the aforementioned tools.The following list of arguments are recognized:

  • the size (font, height)
  • brightness
  • contrast
  • gamma and dither
  • format type of out the output pic

Anyways I found that the basic just in / out arguments passed are enough to produce pretty good results:

hipo@noah:~$ img2txt hipo_avatar.gif >hipo_avatar_pic.txt

The original hipo_avatar.gif file looks like so:

hipo avatar gif picture before img2txt convertion to text

After above img2txt command is run and hipo_avatar_pic.txt to see the colorful output ASCII art img2txt produces, cat it:

hipo@noah:~$ cat hipo_avatar_pic.txt

The image result if screenshot looks quite beautiful and even, can be considered or used as an ART effect image (filter) 🙂

Console Screenshot hipo avatar pic ASCII img2txt output picture

The picture colors are plain ANSI color, so in order to display properly the picture with colors on another computers or Operating System you will need at least basic support for ANSI colors.

Plenty of output file formats are supported by img2txt

Here is the complete list of supported output formats:
 

ansi : coloured ANSI
caca : internal libcaca format
utf8 : UTF8 with CR
utf8 : UTF8 with CRLF (MS Windows)
html : HTML with CSS and DIV support
html3 : Pure HTML3 with tables
irc : IRC with ctrl-k codes
bbfr : BBCode (French)
ps : Postscript
svg : Scalable Vector Graphics
tga : Targa Image

libcaca is available for FreeBSD too, but the caca-utils is not available as a port yet, though probably the deb or rpm packages can easily be ported to BSD.

Well that's all, Enjoy.

How to mount NTFS Windows XP filesystem on FreeBSD, NetBSD, OpenBSD

Friday, May 11th, 2012

Mounting NTFS hdd partitions on FreeBSD logo picture

A friend of mine bring home a Seagate External Hard Disk Drive using USB 3 as a communication media. I needed to attach the hard disk to my FreeBSD router to transfer him some data, the External HDD is formatted to use NTFS as a main file partition and hence to make the file transfers I had to somehow mount the NTFS partition on the HDD.

FreeBSD and other BSDs, just like Linux does not have embedded NTFS file system mount support.
In order to add an external write support for the plugged hdd NTFS I looked in the ports tree:

freebsd# cd /usr/ports
freebsd# make search name='ntfs'
Port: fusefs-ntfs-2010.10.2
Path: /usr/ports/sysutils/fusefs-ntfs
Info: Mount NTFS partitions (read/write) and disk images
Maint: ports@FreeBSD.org
B-deps: fusefs-libs-2.7.4 libiconv-1.13.1_1 libtool-2.4 libublio-20070103 pkg-config-0.25_1
R-deps: fusefs-kmod-0.3.9.p1.20080208_7 fusefs-libs-2.7.4 libiconv-1.13.1_1 libublio-20070103 pkg-config-0.25_1
WWW: http://www.tuxera.com/community/

Port: ntfsprogs-2.0.0_1
Path: /usr/ports/sysutils/ntfsprogs
Info: Utilities and library to manipulate NTFS partitions
Maint: ports@FreeBSD.org
B-deps: fusefs-libs-2.7.4 libiconv-1.13.1_1 libublio-20070103 pkg-config-0.25_1
R-deps: libublio-20070103 pkg-config-0.25_1
WWW: http://www.linux-ntfs.org/
freebs# cd sysutils/fusefs-ntfs/
freebsd# ls
Makefile distinfo files/ pkg-descr pkg-plist
freebsd# cat pkg-descr
The ntfs-3g driver is an open source, freely available read/write NTFS
driver, which provides safe and fast handling of the Windows XP, Windows
Server 2003 and Windows 2000 filesystems. Almost the full POSIX filesystem
functionality is supported, the major exceptions are changing the file
ownerships and the access rights.
WWW: http://www.tuxera.com/community/

Using ntfs-3g I managed to succeed mounting the NTFS on my old PC running FreeBSD ver. 7_2

1. Installing fuserfs-ntfs support on BSD

Before I can use ntfs-3g, to mount the paritition, I had to install fuserfs-ntfs bsd port, with:

freebsd# cd /usr/ports/sysutils/fusefs-ntfs
freebsd# make install clean
.....

I was curious if ntfsprogs provides other utilities to do the ntfs mount but whilst trying to install it I realized it is already installed as a dependency package to fusefs-ntfs.

fusefs-ntfs package provides a number of utilities for creating, mounting, fixing and doing various manipulations with Microsoft NTFS filesystems.

Here is a list of all the executable utilities helpful in NTFS fs management:

freebsd# pkg_info -L fusefs-ntfs\* | grep -E "/bin/|/sbin|README"
/usr/local/bin/lowntfs-3g
/usr/local/bin/ntfs-3g
/usr/local/bin/ntfs-3g.probe
/usr/local/bin/ntfs-3g.secaudit
/usr/local/bin/ntfs-3g.usermap
/usr/local/bin/ntfscat
/usr/local/bin/ntfscluster
/usr/local/bin/ntfscmp
/usr/local/bin/ntfsfix
/usr/local/bin/ntfsinfo
/usr/local/bin/ntfsls
/usr/local/sbin/mkntfs
/usr/local/sbin/ntfsclone
/usr/local/sbin/ntfscp
/usr/local/sbin/ntfslabel
/usr/local/sbin/ntfsresize
/usr/local/sbin/ntfsundelete
/usr/local/share/doc/ntfs-3g/README
/usr/local/share/doc/ntfs-3g/README.FreeBSD

The README and README.FreeBSD are wonderful, reading for those who want to get more in depth knowledge on using the up-listed utilities.

One utility, worthy to mention, I have used in the past is ntfsfix. ntfsfix resolve issues with NTFS partitions which were not unmounted on system shutdown (electricity outage), system hang up etc.

2. Start fusefs (ntfs) and configure it to auto load on system boot

Once fuserfs-ntfs is installed, if its necessery ntfs fs mounts to be permanently supported on the BSD system add fusefs_enable="YES" to /etc/rc.conf(the FreeBSD services auto load conf).

freebsd# echo 'fusefs_enable="YES"' >> /etc/rc.conf

One note to make here is that you need to have also dbus_enable="YES" and hald_enable="YES" in /etc/rc.conf, not having this two in rc.conf will prevent fusefs to start properly. Do a quick grep to make sure this two variables are enabled:

Afterwards fsusefs load up script should be run:

freebsd# /usr/local/etc/rc.d/fusefs start
Starting fusefs.

Another alternative way to load ntfs support on the BSD host is to directly load fuse.ko kernel module:

freebsd# /sbin/kldload fuse.ko

3. Mounting the NTFS partition

In my case, the Seagate hard drive was detected as da0, where the NTFS partition was detected as s1 (da0s1):

freebsd# dmesg|grep -i da0
da0 at umass-sim0 bus 0 target 0 lun 0
da0: Fixed Direct Access SCSI-4 device
da0: 40.000MB/s transfers
da0: 953869MB (1953525164 512 byte sectors: 255H 63S/T 121601C)br />GEOM_LABEL: Label for provider da0s1 is ntfs/Expansion Drive.
GEOM_LABEL: Label for provider da0s1 is ntfs/Expansion Drive.

Therefore further to mount it one can use mount_ntfs (to quickly mount in read only mode) or ntfs-3g for (read / write mode):

If you need to just quickly mount a disk drive to copy some data and umount it with no need for writting to the NTFS partition do;

freebsd# /sbin/mount_ntfs /dev/ad0s1 /mnt/disk

Note that mount_ntfs command is a native BSD command and have nothing to do with ntfs-3g. Therefore using it to mount NTFS is not the same as mounting it via ntfs-3g cmd

freebsd# /usr/local/bin/ntfs-3g -o rw /dev/da0s1 /mnt/disk/

Something, I've noticed while using ntfs-3g is, it fails to properly exit even when the ntfs-3g shell execution is over:

freebsd# ps ax |grep -i ntfs|grep -v grep
18892 ?? Is 0:00.00 /usr/local/bin/ntfs-3g -o rw /dev/da0s1 /mnt/disk/

I dunno if this is some kind of ntfs-3g bug or feature specific to all versions of FreeBSD or it is something local to FBSD 7.2

Thought ntfs-3g, keeps appearing in process list, praise God as of time of writting NTFS support on FreeBSD prooved to be stable.
Read / Write disk operations to the NTFS I tested it with works great. Just about 5 years ago I still remember write mode was still experimental. Now it seems NTFS mounts can be used with no hassle even on production machines.

4. Auto mounting NTFS partition on FreeBSD system boot

There are two approaches towards 'the problem' I can think of.
The better way to auto mount on boot (in my view) is through /etc/fstab use

If /etc/fstab + ntfs-3g is to be used, you will however change the default /sbin/mount_ntfs command to point to /usr/local/bin/ntfs-3g, i.e.:

freebsd# mv /sbin/mount_ntfs /sbin/mount_ntfs.orig
freebsd# ln -s /usr/local/bin/ntfs-3g /sbin/mount_ntfs

Then to mount /dev/da0s1 via /etc/fstab add line:

/dev/ad0s1 /mnt/disk ntfs rw,late 0 0

To not bother with text editor run:

freebsd# echo '/dev/ad0s1 /mnt/disk ntfs rw,late 0 0' >> /etc/fstab

I've red in posts in freebsd forums, there is also a way to use ntfs-3g for mounting partitions, without substituting the original bsd /sbin/mount_ntfs, the exact commands suggested to be used with no need to prior mv /sbin/mount_ntfs to /sbin/mount_ntfs.orig and link it to ntfs is:

/dev/ad0s1 /disk ntfs rw,mountprog=/usr/local/bin/ntfs-3g,late 0 0

For any other NTFS partitions, for instance /dev/ad0s2, /dev/ad2s1 etc. simply change the parititon name and mount points.

The second alternative to adding the NTFS to auto mount is through /etc/rc.local. /etc/rc.local content will be executed very late in system boot. :

echo '/usr/local/bin/ntfs-3g -o rw /dev/da0s1' >> /etc/rc.local

One disadvanage of using /etc/rc.local for mounting the partition is the hanging ntfs-3g in proc list:

freebsd# ps ax |grep -i ntfs|grep -v grep
18892 ?? Is 0:00.00 /usr/local/bin/ntfs-3g -o rw /dev/da0s1 /mnt/disk/

Though, I haven't tested it yet. Using the same methodology should be perfectly working on PC-BSD, DragonFlyBSD, NetBSD and OpenBSD.
I will be glad if someone who runs any of the other BSDs can confirm, following this instructions works fine on these BSDs too.

How to solve “Incorrect key file for table ‘/tmp/#sql_9315.MYI’; try to repair it” mysql start up error

Saturday, April 28th, 2012

When a server hard disk scape gets filled its common that Apache returns empty (no content) pages…
This just happened in one server I administer. To restore the normal server operation I freed some space by deleting old obsolete backups.
Actually the whole reasons for this mess was an enormous backup files, which on the last monthly backup overfilled the disk empty space.

Though, I freed about 400GB of space on the the root filesystem and on a first glimpse the system had plenty of free hard drive space, still restarting the MySQL server refused to start up properly and spit error:

Incorrect key file for table '/tmp/#sql_9315.MYI'; try to repair it" mysql start up error

Besides that there have been corrupted (crashed) tables, which reported next to above error.
Checking in /tmp/#sql_9315.MYI, I couldn't see any MYI – (MyISAM) format file. A quick google look up revealed that this error is caused by not enough disk space. This was puzzling as I can see both /var and / partitions had plenty of space so this shouldn't be a problem. Also manally creating the file /tmp/#sql_9315.MYI with:

server:~# touch /tmp/#sql_9315.MYI

Didn't help it, though the file created fine. Anyways a bit of a closer examination I've noticed a /tmp filesystem mounted besides with the other file system mounts ????
You can guess my great amazement to find this 1 Megabyte only /tmp filesystem hanging on the server mounted on the server.

I didn't mounted this 1 Megabyte filesystem, so it was either an intruder or some kind of "weird" bug…
I digged in Googling to see, if I can find more on the error and found actually the whole mess with this 1 mb mounted /tmp partition is caused by, just recently introduced Debian init script /etc/init.d/mountoverflowtmp.
It seems this script was introduced in Debian newer releases. mountoverflowtmp is some kind of emergency script, which is triggered in case if the root filesystem/ space gets filled.
The script has only two options:

# /etc/init.d/mountoverflowtmp
Usage: mountoverflowtmp [start|stop]

Once started what it does it remounts the /tmp to be 1 megabyte in size and stops its execution like it never run. Well maybe, the developers had something in mind with introducing this script I will not argue. What I should complain though is the script design is completely broken. Once the script gets "activated" and does its job. This 1MB mount stays like this, even if hard disk space is freed on the root partition – / ….

Hence to cope with this unhandy situation, once I had freed disk space on the root partition for some reason mountoverflowtmp stop option was not working,
So I had to initiate "hard" unmount:

server:~# mount -l /tmp

Also as I had a bunch of crashed tables and to fix them, also issued on each of the broken tables reported on /etc/init.d/mysql start start-up.

server:~# mysql -u root -p
mysql> use Database_Name;
mysql> repair table Table_Name extended;
....

Then to finally solve the stupid Incorrect key file for table '/tmp/#sql_XXYYZZ33444.MYI'; try to repair it error, I had to restart once again the SQL server:

Stopping MySQL database server: mysqld.
Starting MySQL database server: mysqld.
Checking for corrupt, not cleanly closed and upgrade needing tables..
root@server:/etc/init.d#

Tadadadadam!, SQL now loads and works back as before!

Tracking I/O hard disk server bottlenecks with iostat on GNU / Linux and FreeBSD

Tuesday, March 27th, 2012

Hard disk overhead tracking on Linux and FreeBSD with iostat

I've earlier wrote an article How to find which processes are causing hard disk i/o overhead on Linux there I explained very rawly few tools which can be used to benchmark hard disk read / write operations. My prior article accent was on iotop and dstat and it just mentioned of iostat. Therefore I've wrote this short article in attempt to explain a bit more thoroughfully on how iostat can be used to track problems with excessive server I/O read/writes.

Here is the command man page description;
iostatReport Central Processing Unit (CPU) statistics and input/output statistics for devices, partitions and network filesystems

I will further proceed with few words on how iostat can be installed on various Linux distros, then point at few most common scenarious of use and a short explanation on the meaning of each of the command outputs.

1. Installing iostat on Linux

iostat is a swiss army knife of finding a server hard disk bottlenecks. Though it is a must have tool in the admin outfut, most of Linux distributions will not have iostat installed by default.
To have it on your server, you will need to install sysstat package:

a) On Debian / Ubuntu and other Debian GNU / Linux derivatives to install sysstat:

debian:~# apt-get --yes install sysstat

b) On Fedora, CentOS, RHEL etc. install is with yum:

[root@centos ~]# yum -y install sysstat

c) On Slackware Linux sysstat package which contains iostat is installed by default. 

d) In FreeBSD, there is no need for installation of any external package as iostat is part of the BSD world (bundle commands).
I should mention bsd iostat and Linux's iostat commands are not the same and hence there use to track down hard disk bottlenecks differs a bit, however the general logic of use is very similar as with most tools in BSD and Linux.

2. Checking a server hard disk for i/o disk bottlenecks on G* / Linux

Once having the sysstat installed on G* / Linux systems, the iostat command will be added in /usr/bin/iostat
a) To check what is the hard disk read writes per second (in megabytes) use:

debian:~# /usr/bin/iostat -m
Linux 2.6.32-5-amd64 (debian) 03/27/2012 _x86_64_ (8 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
15.34 0.36 2.76 2.66 0.00 78.88
Device: tps MB_read/s MB_wrtn/s MB_read MB_wrtn
sda 63.89 0.48 8.20 6730223 115541235
sdb 64.12 0.44 8.23 6244683 116039483
md0 2118.70 0.22 8.19 3041643 115528074

In the above output the server, where I issue the command is using sda and sdb configured in software RAID 1 array visible in the output as (md0)

The output of iostat should already be easily to read, for anyone who didn't used the tool here is a few lines explanation of the columns:

The %user 15.34 meaning is that 15.34 out of 100% possible i/o load is generad by system level read/write operations.
%nice – >Show the percentage of CPU utilization that occurred while executing at the user level with nice priority.
%iowait – just like the top command idle it shows the idle time when the system didn't have an outstanding disk I/O requests.
%steal – show percentage in time spent in time wait of CPU or virtual CPUs to service another virtual processor (high numbers of disk is sure sign for i/o problem).
%idle – almost the same as meaning to %iowait
tps – HDD transactions per second
MB_read/s (column) – shows the actual Disk reads in Mbytes at the time of issuing iostat
MB_wrtn/s – displays the writes p/s at the time of iostat invocation
MB_read – shows the hard disk read operations in megabytes, since the server boot 'till moment of invocation of iostat
MB_wrtn – gives the number of Megabytes written on HDD since the last server boot filesystem mount

The reason why the Read / Write values for sda and sdb are similar in this example output is because my disks are configured in software RAID1 (mirror)

The above iostat output reveals in my specific case the server is experiencing mostly Disk writes (observable in the high MB_wrtn/s 8.19 md0 in the above sample output).

It also reveals, the I/O reads experienced on that server hard disk are mostly generated as a system (user level load) – see (%user 15.34 and md0 2118.70).

For all those not familiar with system also called user / level load, this is all kind of load which is generated by running programs on the server – (any kind of load not generated by the Linux kernel or loaded kernel modules).

b) To periodically keep an eye on HDD i/o operations with iostat, there are two ways:

– Use watch in conjunction with iostat;

[root@centos ~]# watch "/usr/bin/iostat -m"
Every 2.0s: iostat -m Tue Mar 27 11:00:30 2012
Linux 2.6.32-5-amd64 (centos) 03/27/2012 _x86_64_ (8 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
15.34 0.36 2.76 2.66 0.00 78.88
Device: tps MB_read/s MB_wrtn/s MB_read MB_wrtn
sda 63.89 0.48 8.20 6730255 115574152
sdb 64.12 0.44 8.23 6244718 116072400
md0 2118.94 0.22 8.20 3041710 115560990
Device: tps MB_read/s MB_wrtn/s MB_read MB_wrtn
sda 55.00 0.01 25.75 0 51
sdb 52.50 0.00 24.75 0 49
md0 34661.00 0.01 135.38 0 270

Even though watch use and -d might appear like identical, they're not watch does refresh the screen, executing instruction similar to the clear command which clears screen on every 2 seconds, so the output looks like the top command refresh, while passing the -d 2 will output the iostat command output on every 2 secs in a row so all the data is visualized on the screen. Hence -d 2 in cases, where more thorough debug is necessery is better. However for a quick routine view watch + iostat is great too.

c) Outputting extra information for HDD input/output operations;

root@debian:~# iostat -x
Linux 2.6.32-5-amd64 (debian) 03/27/2012 _x86_64_ (8 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
15.34 0.36 2.76 2.66 0.00 78.88
Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s avgrq-sz avgqu-sz await svctm %util
sda 4.22 2047.33 12.01 51.88 977.44 16785.96 278.03 0.28 4.35 3.87 24.72
sdb 3.80 2047.61 11.97 52.15 906.93 16858.32 277.05 0.03 5.25 3.87 24.84
md0 0.00 0.00 20.72 2098.28 441.75 16784.05 8.13 0.00 0.00 0.00 0.00

This command will output extended useful Hard Disk info like;
r/s – number of read requests issued per second
w/s – number of write requests issued per second
rsec/s – numbers of sector reads per second
b>wsec/s – number of sectors wrote per second
etc. etc.

Most of ppl will never need to use this, but it is good to know it exists.

3. Tracking read / write (i/o) hard disk bottlenecks on FreeBSD

BSD's iostat is a bit different in terms of output and arguments.

a) Here is most basic use:

freebsd# /usr/sbin/iostat
tty ad0 cpu
tin tout KB/t tps MB/s us ni sy in id
1 561 45.18 44 1.95 14 0 5 0 82

b) Periodic watch of hdd i/o operations;

freebsd# iostat -c 10
tty ad0 cpu
tin tout KB/t tps MB/s us ni sy in id
1 562 45.19 44 1.95 14 0 5 0 82
0 307 51.96 113 5.73 44 0 24 0 32
0 234 58.12 98 5.56 16 0 7 0 77
0 43 0.00 0 0.00 1 0 0 0 99
0 485 0.00 0 0.00 2 0 0 0 98
0 43 0.00 0 0.00 0 0 1 0 99
0 43 0.00 0 0.00 0 0 0 0 100
...

As you see in the output, there is information like in the columns tty, tin, tout which is a bit hard to comprehend.
Thanksfully the tool has an option to print out only more essential i/o information:

freebsd# iostat -d -c 10
ad0
KB/t tps MB/s
45.19 44 1.95
58.12 97 5.52
54.81 108 5.78
0.00 0 0.00
0.00 0 0.00
0.00 0 0.00
20.48 25 0.50

The output info is quite self-explanatory.

Displaying a number of iostat values for hard disk reads can be also achieved by omitting -c option with:

freebsd# iostat -d 1 10
...

Tracking a specific hard disk partiotion with iostat is done with:

freebsd# iostat -n /dev/ad0s1a
tty cpu
tin tout us ni sy in id
1 577 14 0 5 0 81
c) Getting Hard disk read/write information with gstat

gstat is a FreeBSD tool to print statistics for GEOM disks. Its default behaviour is to refresh the screen in a similar fashion like top command, so its great for people who would like to periodically check all attached system hard disk and storage devices:

freebsd# gstat
dT: 1.002s w: 1.000s
L(q) ops/s r/s kBps ms/r w/s kBps ms/w %busy Name
0 10 0 0 0.0 10 260 2.6 15.6| ad0
0 10 0 0 0.0 10 260 2.6 11.4| ad0s1
0 10 0 0 0.0 10 260 2.8 12.5| ad0s1a
0 0 0 0 0.0 0 0 0.0 20.0| ad0s1b
0 0 0 0 0.0 0 0 0.0 0.0| ad0s1c
0 0 0 0 0.0 0 0 0.0 0.0| ad0s1d
0 0 0 0 0.0 0 0 0.0 0.0| ad0s1e
0 0 0 0 0.0 0 0 0.0 0.0| acd0

It even has colors if your tty supports colors 🙂

Another useful tool in debugging the culprit of excessive hdd I/O operations is procstat command:

Here is a sample procstat run to track (httpd) one of my processes imposing i/o hdd load:

freebsd# procstat -f 50404
PID COMM FD T V FLAGS REF OFFSET PRO NAME
50404 httpd cwd v d -------- - - - /
50404 httpd root v d -------- - - - /
50404 httpd 0 v c r------- 56 0 - -
50404 httpd 1 v c -w------ 56 0 - -
50404 httpd 2 v r -wa----- 56 75581 - /var/log/httpd-error.log
50404 httpd 3 s - rw------ 105 0 TCP ::.80 ::.0
50404 httpd 4 p - rw---n-- 56 0 - -
50404 httpd 5 p - rw------ 56 0 - -
50404 httpd 6 v r -wa----- 56 25161132 - /var/log/httpd-access.log
50404 httpd 7 v r rw------ 56 0 - /tmp/apr8QUOUW
50404 httpd 8 v r -w------ 56 0 - /var/run/accept.lock.49588
50404 httpd 9 v r -w------ 1 0 - /var/run/accept.lock.49588
50404 httpd 10 v r -w------ 1 0 - /tmp/apr8QUOUW
50404 httpd 11 ? - -------- 2 0 - -

Btw fstat is sometimes helpful in identifying the number of open files and trying to estimate which ones are putting the hdd load.
Hope this info helps someone. If you know better ways to track hdd excessive loads on Linux / BSD pls share 'em pls.
 

How to delete million of files on busy Linux servers (Work out Argument list too long)

Tuesday, March 20th, 2012

How to Delete million or many thousands of files in the same directory on GNU / Linux and FreeBSD

If you try to delete more than 131072 of files on Linux with rm -f *, where the files are all stored in the same directory, you will get an error:

/bin/rm: Argument list too long.

I've earlier blogged on deleting multiple files on Linux and FreeBSD and this is not my first time facing this error.
Anyways, as time passed, I've found few other new ways to delete large multitudes of files from a server.

In this article, I will explain shortly few approaches to delete few million of obsolete files to clean some space on your server.
Here are 3 methods to use to clean your tons of junk files.

1. Using Linux find command to wipe out millions of files

a.) Finding and deleting files using find's -exec switch:

# find . -type f -exec rm -fv {} \;

This method works fine but it has 1 downside, file deletion is too slow as for each found file external rm command is invoked.

For half a million of files or more, using this method will take "long". However from a server hard disk stressing point of view it is not so bad as, the files deletion is not putting too much strain on the server hard disk.
b.) Finding and deleting big number of files with find's -delete argument:

Luckily, there is a better way to delete the files, by using find's command embedded -delete argument:

# find . -type f -print -delete

c.) Deleting and printing out deleted files with find's -print arg

If you would like to output on your terminal, what files find is deleting in "real time" add -print:

# find . -type f -print -delete

To prevent your server hard disk from being stressed and hence save your self from server normal operation "outages", it is good to combine find command with ionice, e.g.:

# ionice -c 3 find . -type f -print -delete

Just note, that ionice cannot guarantee find's opeartions will not affect severely hard disk i/o requests. On  heavily busy servers with high amounts of disk i/o writes still applying the ionice will not prevent the server from being hanged! Be sure to always keep an eye on the server, while deleting the files nomatter with or without ionice. if throughout find execution, the server gets lagged in serving its ordinary client requests or whatever, stop the execution of the cmd immediately by killing it from another ssh session or tty (if physically on the server).

2. Using a simple bash loop with rm command to delete "tons" of files

An alternative way is to use a bash loop, to print each of the files in the directory and issue /bin/rm on each of the loop elements (files) like so:

for i in *; do
rm -f $i;
done

If you'd like to print what you will be deleting add an echo to the loop:

# for i in $(echo *); do \
echo "Deleting : $i"; rm -f $i; \

The bash loop, worked like a charm in my case so I really warmly recommend this method, whenever you need to delete more than 500 000+ files in a directory.

3. Deleting multiple files with perl

Deleting multiple files with perl is not a bad idea at all.
Here is a perl one liner, to delete all files contained within a directory:

# perl -e 'for(<*>){((stat)[9]<(unlink))}'

If you prefer to use more human readable perl script to delete a multitide of files use delete_multple_files_in_dir_perl.pl

Using perl interpreter to delete thousand of files is quick, really, really quick.
I did not benchmark it on the server, how quick exactly is it, but I guess the delete rate should be similar to find command. Its possible even in some cases the perl loop is  quicker …

4. Using PHP script to delete a multiple files

Using a short php script to delete files file by file in a loop similar to above bash script is another option.
To do deletion  with PHP, use this little PHP script:

<?php
$dir = "/path/to/dir/with/files";
$dh = opendir( $dir);
$i = 0;
while (($file = readdir($dh)) !== false) {
$file = "$dir/$file";
if (is_file( $file)) {
unlink( $file);
if (!(++$i % 1000)) {
echo "$i files removed\n";
}
}
}
?>

As you see the script reads the $dir defined directory and loops through it, opening file by file and doing a delete for each of its loop elements.
You should already know PHP is slow, so this method is only useful if you have to delete many thousands of files on a shared hosting server with no (ssh) shell access.

This php script is taken from Steve Kamerman's blog . I would like also to express my big gratitude to Steve for writting such a wonderful post. His post actually become  inspiration for this article to become reality.

You can also download the php delete million of files script sample here

To use it rename delete_millioon_of_files_in_a_dir.php.txt to delete_millioon_of_files_in_a_dir.php and run it through a browser .

Note that you might need to run it multiple times, cause many shared hosting servers are configured to exit a php script which keeps running for too long.
Alternatively the script can be run through shell with PHP cli:

php -l delete_millioon_of_files_in_a_dir.php.txt.

5. So What is the "best" way to delete million of files on Linux?

In order to find out which method is quicker in terms of execution time I did a home brew benchmarking on my thinkpad notebook.

a) Creating 509072 of sample files.

Again, I used bash loop to create many thousands of files in order to benchmark.
I didn't wanted to put this load on a productive server and hence I used my own notebook to conduct the benchmarks. As my notebook is not a server the benchmarks might be partially incorrect, however I believe still .they're pretty good indicator on which deletion method would be better.

hipo@noah:~$ mkdir /tmp/test
hipo@noah:~$ cd /tmp/test;
hiponoah:/tmp/test$ for i in $(seq 1 509072); do echo aaaa >> $i.txt; done

I had to wait few minutes until I have at hand 509072  of files created. Each of the files as you can read is containing the sample "aaaa" string.

b) Calculating the number of files in the directory

Once the command was completed to make sure all the 509072 were existing, I used a find + wc cmd to calculate the directory contained number of files:

hipo@noah:/tmp/test$ find . -maxdepth 1 -type f |wc -l
509072

real 0m1.886s
user 0m0.440s
sys 0m1.332s

Its intesrsting, using an ls command to calculate the files is less efficient than using find:

hipo@noah:/tmp/test$ time ls -1 |wc -l
509072

real 0m3.355s
user 0m2.696s
sys 0m0.528s

c) benchmarking the different file deleting methods with time

– Testing delete speed of find

hipo@noah:/tmp/test$ time find . -maxdepth 1 -type f -delete
real 15m40.853s
user 0m0.908s
sys 0m22.357s

You see, using find to delete the files is not either too slow nor light quick.

– How fast is perl loop in multitude file deletion ?

hipo@noah:/tmp/test$ time perl -e 'for(<*>){((stat)[9]<(unlink))}'real 6m24.669suser 0m2.980ssys 0m22.673s

Deleting my sample 509072 took 6 mins and 24 secs. This is about 3 times faster than find! GO-GO perl 🙂
As you can see from the results, perl is a great and time saving, way to delete 500 000 files.

– The approximate speed deletion rate of of for + rm bash loop

hipo@noah:/tmp/test$ time for i in *; do rm -f $i; done

real 206m15.081s
user 2m38.954s
sys 195m38.182s

You see the execution took 195m en 38 secs = 3 HOURS and 43 MINUTES!!!! This is extremely slow ! But works like a charm as the running of deletion didn't impacted my normal laptop browsing. While the script was running I was mostly browsing through few not so heavy (non flash) websites and doing some other stuff in gnome-terminal) 🙂

As you can imagine running a bash loop is a bit CPU intensive, but puts less stress on the hard disk read/write operations. Therefore its clear using it is always a good practice when deletion of many files on a dedi servers is required.

b) my production server file deleting experience

On a production server I only tested two of all the listed methods to delete my files. The production server, where I tested is running Debian GNU / Linux Squeeze 6.0.3. There I had a task to delete few million of files.
The tested methods tried on the server were:

– The find . type -f -delete method.

– for i in *; do rm -f $i; done

The results from using find -delete method was quite sad, as the server almost hanged under the heavy hard disk load the command produced.

With the for script all went smoothly. The files were deleted for a long long time (like few hours), but while it was running, the server continued with no interruptions..

While the bash loop was running, the server load avarage kept at steady 4
Taking my experience in mind, If you're running a production, server and you're still wondering which delete method to use to wipe some multitude of files, I would recommend you go  the bash for loop + /bin/rm way. Yes, it is extremely slow, expect it run for some half an hour or so but puts not too much extra load on the server..

Using the PHP script will probably be slow and inefficient, if compared to both find and the a bash loop.. I didn't give it a try yet, but suppose it will be either equal in time or at least few times slower than bash.

If you have tried the php script and you have some observations, please drop some comment to tell me how it performs.

To sum it up;

Even though there are "hacks" to clean up some messy parsing directory full of few million of junk files, having such a directory should never exist on the first place.

Frankly, keeping millions of files within the same directory is very stupid idea.
Doing so will have a severe negative impact on a directory listing performance of your filesystem in the long term.

If you know better (more efficient) ways to delete a multitude of files in a dir please share in comments.

diskinfo Linux hdparm FreeBSD equivalent command for disk info and benchmarking

Thursday, March 8th, 2012

FreeBSD Linux hdparm equivalent is diskinfo artistic logo

On Linux there is the hdparm tool for various hard disk benchmarking and extraction of hard disk operations info.
As the Linux manual states hdparmget/set SATA/IDE device parameters

Most Linux users should already know it and might wonder if there is hdparm port or equivalent for FreeBSD, the aim of this short post is to shed some light on that.

The typical use of hdparm is like this:

linux:~# hdparm -t /dev/sda8

/dev/sda8:
Timing buffered disk reads: 76 MB in 3.03 seconds = 25.12 MB/sec
linux:~# hdparm -T /dev/sda8
/dev/sda8:
Timing cached reads: 1618 MB in 2.00 seconds = 809.49 MB/sec

The above output here is from my notebook Lenovo R61i.
If you're looking for alternative command to hdparm you should know in FreeBSD / OpenBSD / NetBSD, there is no exact hdparm equivalent command.
The somehow similar hdparm equivallent command for BSDs (FreeBSD etc.) is:
diskinfo

diskinfo is not so feature rich as linux's hdparm. It is just a simple command to show basic information for hard disk operations without no possibility to tune any hdd I/O and seek operations.
All diskinfo does is to show statistics for a hard drive seek times I/O overheads. The command takes only 3 arguments.

The most basic and classical use of the command is:

freebsd# diskinfo -t /dev/ad0s1a
/dev/ad0s1a
512 # sectorsize
20971520000 # mediasize in bytes (20G)
40960000 # mediasize in sectors
40634 # Cylinders according to firmware.
16 # Heads according to firmware.
63 # Sectors according to firmware.
ad:4JV48BXJs0s0 # Disk ident.

Seek times:
Full stroke: 250 iter in 3.272735 sec = 13.091 msec
Half stroke: 250 iter in 3.507849 sec = 14.031 msec
Quarter stroke: 500 iter in 9.705555 sec = 19.411 msec
Short forward: 400 iter in 2.605652 sec = 6.514 msec
Short backward: 400 iter in 4.333490 sec = 10.834 msec
Seq outer: 2048 iter in 1.150611 sec = 0.562 msec
Seq inner: 2048 iter in 0.215104 sec = 0.105 msec

Transfer rates:
outside: 102400 kbytes in 3.056943 sec = 33498 kbytes/sec
middle: 102400 kbytes in 2.696326 sec = 37978 kbytes/sec
inside: 102400 kbytes in 3.178711 sec = 32214 kbytes/sec

Another common use of diskinfo is to measure hdd I/O command overheads with -c argument:

freebsd# diskinfo -c /dev/ad0s1e
/dev/ad0s1e
512 # sectorsize
39112312320 # mediasize in bytes (36G)
76391235 # mediasize in sectors
75784 # Cylinders according to firmware.
16 # Heads according to firmware.
63 # Sectors according to firmware.
ad:4JV48BXJs0s4 # Disk ident.

I/O command overhead:
time to read 10MB block 1.828021 sec = 0.089 msec/sector
time to read 20480 sectors 4.435214 sec = 0.217 msec/sector
calculated command overhead = 0.127 msec/sector

Above diskinfo output is from my FreeBSD home router.

As you can see, the time to read 10MB block on my hard drive is 1.828021 (which is very high number),
this is a sign the hard disk experience too many read/writes and therefore needs to be shortly replaced with newer faster one.
diskinfo is part of the basis bsd install (bsd world). So it can be used without installing any bsd ports or binary packages.

For the purpose of stress testing hdd, or just some more detailed benchmarking on FreeBSD there are plenty of other tools as well.
Just to name a few:
 

  • rawio – obsolete in FreeBSD 7.x version branch (not available in BSD 7.2 and higher)
  • iozone, iozone21 – Tools to test the speed of sequential I/O to files
  • bonnie++ – benchmark tool capable of performing number of simple fs tests
  • bonnie – predecessor filesystem benchmark tool to bonnie++
  • raidtest – test performance of storage devices
  • mdtest – Software to test metadata performance on filsystems
  • filebench – tool for micro-benchmarking storage subsystems

Linux hdparm allows also changing / setting various hdd ATA and SATA settings. Similarly, to set and change ATA / SATA settings on FreeBSD there is the:

  • ataidle

tool.

As of time of writting ataidle is in port path /usr/ports/sysutils/ataidle/

To check it out install it as usual from the port location:

FreeBSD also has also the spindown port – a small program for handling automated spinning down ofSCSI harddrive
spindown is useful in setting values to SATA drives which has problems with properly controlling HDD power management.

To keep constant track on hard disk operations and preliminary warning in case of failing hard disks on FreeBSD there is also smartd service, just like in Linux.
smartd enables you to to control and monitor storage systems using the Self-Monitoring, Analysisand Reporting Technology System (S.M.A.R.T.) built into most modern ATA and SCSI hard disks.
smartd and smartctl are installable via the port /usr/ports/sysutils/smartmontools.

To install and use smartd, ataidle and spindown run:

freebsd# cd /usr/ports/sysutils/smartmontools
freebsd# make && make install clean
freebsd# cd /usr/ports/sysutils/ataidle/
freebsd# make && make install clean
freebsd# cd /usr/ports/sysutils/spindown/
freebsd# make && make install clean

Check each one's manual for more info.

Recommended logrorate practices on heavy loaded (busy) Apache Linux servers

Wednesday, March 7th, 2012

Apache logrotate Debian good configuration for heavy loaded servers

If you are sys admin of Apache Webserver running on Debian Linux relying on logrorate to rorate logs, you might want to change the default way logroration is done.

Little changes in the way Apache log files are served on busy servers can have positive outcomes on the overall way the server CPU units burden. A good logrotation strategy can also prevent your server from occasional extra overheads or downtimes.

The way Debian GNU / Linux process logs is well planned for small servers, however the default logroration Apache routine doesn't fit well for servers which process millions of client requests each day.

I happen to administrate, few servers which are constantly under a heavy load and have occasionally overload troubles because of Debian's logrorate default mechanism.

To cope with the situation I have made few modifications to /etc/logrorate.d/apache2 and decided to share it here hoping, this might help you too.

1. Rotate Apache acccess.log log file daily instead of weekly

On Debian Apache's logrorate script is in /etc/logrotate.d/apache2

The default file content will be like so like so:

debian:~# cat /etc/logrotate.d/apache2
/var/log/apache2/*.log {
weekly
missingok
rotate 52
size 1G
compress
delaycompress
notifempty
create 640 root adm
sharedscripts
postrotate
if [ -f "`. /etc/apache2/envvars ; echo ${APACHE_PID_FILE:-/var/run/apache2.pid}`" ]; then
/etc/init.d/apache2 reload > /dev/null
fi
endscript
}

To change the rotation from weekly to daily change:

weekly

to

#weekly

2. Disable access.log log file gzip compression

By default apache2 logrotate script is tuned ot make compression of rotated file (exmpl: copy access.log to access.log.1 and gzip it, copy access.log to access.log.2 and gzip it etc.). On servers where logs are many gigabytes, once logrotate initiates its scheduled work it will have to compress an enormous log record of apache requests. On very busy Apache servers from my experience, just for a day the log could grow up to approximately 8 / 10 Gigabytes.
I'm sure there are more busy servers out there, which log files are growing to over 100GB for just a single day.
Gzipping a 100GB file piece takes an enormous load on the CPU, as well as often takes long time. When this logrotation gzipping occurs at a moment where the servers CPU cores are already heavy loaded from Apache serving HTTP requests, Apache server becomes inaccessible to most of the clients.
Then for end clients various oddities are experienced, for example Apache dropped connection errors, webserver returning empty pages, or simply inability to respond to the client browser.
Sometimes as a result of the overload, even secure shell connection to SSHD to the server is impossible …

To prevent your server from this roration overloads remove logrorate's default access.log gzipping by commenting:

compress

to

#comment

3. Change maximum log roration by logrorate to be up to 30

By default logrorate is configured to create and keep up to 52 rotated and gzipped access.log files, changing this to a lower number is a good practice (in my view), in cases where log files grow daily to 10 or more GBs. Doing so will save a lot of disk space and reduce the chance the hard disk gets filled in because of the multiple rorated ungzipped enormous access.log files.

To tune the default keep max rorated logs to 30, change:

rotate 52

to rotate 30

The way logrorate's apache log processing on RHEL / CentOS Linux is working better on high load servers, by default on CentOS logrorate is not configured to do log gzipping at all.

Here is the default /etc/logrorate.d/httpd script for
CentOS release 5.6 (Final)

[hipo@centos httpd]$ cat /etc/logrotate.d/httpd /var/log/httpd/*log {
missingok
notifempty
sharedscripts
postrotate
/bin/kill -HUP `cat /var/run/httpd.pid 2>/dev/null` 2> /dev/null || true
endscript
}