Unix/Linux

From dbawiki
Revision as of 00:05, 6 December 2018 by Stuart (talk | contribs) (→‎Use sed to repeat characters)
Jump to: navigation, search

Contents

How to automatically download and process email attachments from Gmail

Useful references

Eventually, after months of trying different things, came up with this combination of tools that worked very well.

Install tools

sudo dnf -y install fetchmail maildrop uudeview openssl
  • Set Gmail account up for POP and disable IMAP

Configure fetchmail

This is the part that talks to Gmail and downloads the emails to the local machine. 

vi ~/.fetchmailrc

set postmaster "<username>"
#set daemon 600
poll pop.gmail.com with proto POP3
    user "[email protected]" there with password '<password>' is <username> here options ssl nokeep
    mda "/usr/bin/maildrop .mailfilter"
#    sslcertfile /etc/ssl/certs/ca-bundle.crt

If fetchmail complains about ssl certificates, try specifying wherein is with the sslcertfile line

Configure maildrop

vi ~/.mailfilter

LOGFILE = "/home/dbahawk/maildrop/maildrop.log"
DEFAULT="$HOME/maildrop"

# keep a copy
cc backup
`cd backup/new && rm -f dummy \`ls -t | sed -e 1,50d\``

if (/^To:.*getmail@...\.dnsalias\.org$/)
   {
        to Maildir/getmail/
   }

#if (/^To:.*dbahawk/)
#   {
#        to maildrop
#   }

if (/^To:.*dbahawk/)
    dotlock "auto.lock" {
    to "|uudeview -c -i"
}

Not really sure why this filter is not working properly. The cc works but the To: check doesn't. So I have a shell that runs separately to process the emails in the backup mail directory.

uudeview

Use sed to repeat characters

echo an '=' sign then use the loop feature of sed to run round a loop a further 79 times replacing what you've got with the same thing plus and extra one! 

echo "=" | sed -e :a -e 's/^=\{1,79\}$/&=/;ta'

Convert between lower and upper case using ansi standard

This way should be portable and give predictable results 

LOWER_SID=$(echo "${SID}" | tr '[:upper:]' '[:lower:]')
UPPER_SID=$(echo "${SID}" | tr '[:lower:]' '[:upper:]')

Use sed to delete the first line of a file or command output printing from the second line onwards

Actually deceptively simple. Example remove the header from df output. 

df -k | sed 1d

As a useful extension, suppose TO_DATA_DIR contains multiple directories, this will total the free space for all relevant filesystems
This one is specific to AIX, for other Unixes and Linux, use $4 instead of $3. 

SPACE_AVAILABLE=$(df -k $TO_DATA_DIR | sed 1d | awk 'BEGIN {i=0} {i=i+$3} END {print i}')

Use sed to munge a controlfile into submission!

  • delete comment lines
  • delete all lines after the one starting with a semicolon
  • delete all blank lines
  • change reuse to set
  • change archivelog (with word boundaries) to noarchivelog
  • change old sid to new sid
sqlplus -s / as sysdba<<'EOSQL' >/dev/null
alter database backup controlfile to trace as '/tmp/regenerate_controlfile.sql' reuse resetlogs;
EOSQL
OLDSID='PROD'
NEWSID='TEST'
sed  -e '/^--/d' -e '/^\;/q' -e '/^ *$/d' -e 's/REUSE/SET/' -e 's/\<ARCHIVELOG\>/NOARCHIVELOG/' -e 's/\"'${OLDSID}'\"/\"'${NEWSID}'\"/' /tmp/regenerate_controlfile.sql

Find all files where a certain text does not exist

Mostly we want to search for files containing a particular string but how to find those files that do not contain some text.
Using xargs instead of -exec forces the filename to be printed. -H should do this but doesn't seem to in this case. 

find . -name "*cfg" | xargs grep -H -c TO_CV_HOST {} \; | grep ':0$' | cut -d: -f1

Use tar and gzip to copy files more efficiently across network between hosts

From the destination server, this will connect to the source, tar up each file and pull it back to the current server 

ssh oracle@hn5211 "cd /oracle/ora_data4/iten3/ && tar -cf - . | gzip " | ( cd /oracle/ora_data2/iten/ && gunzip -c | tar -xvf - . )

double hash and double percent in shell variables to trim off characters from variables

# - trims everything from the start of line till the FIRST occurrence of following character (abstemious) :-)
## - trims everything from the start of line till the LAST occurrence of following character (greedy)
% - trims everything from the end of line backwards till the FIRST occurrence of following character (abstemious)
%% - trims everything from the end of line backwards till the LAST occurrence of following character (greedy) 

FILENAME="/home/bey9at77/hello.txt"
FILE_STUB1=${FILENAME##*/}
FILE_STUB=${FILE_STUB1%.*}
echo $FILE_STUB
hello
FILE_EXT=${FILENAME##*.}
echo $FILE_EXT
txt

or
# - remove prefix reluctantly
## - remove prefix greedily
% - remove suffix reluctantly
%% - remove suffix greedily

words="do.re.mi"

echo ${words#*.}

echo ${words##*.}

echo ${words%.*}

echo ${words%%.*}

The best shell script ever. An example of how scripting should be done

It was written by someone at Oracle. Unfortunately (s)he did not put any author comment in it. Also unfortunately I cannot show it here as it is protected behind Oracles support website.
If you have an Oracle Metalink id, you can get the complete script here
Here is a snippet of a function that demonstrates proper commenting and a very good style.
The script is called physru.sh and upgrades an Oracle database in a rolling upgrade fashion by using a physical standby.
There are 4500 lines in the full script but it is so easy to read and understand because of the way it's written, it's like a breath of fresh air. Well done whoever you are! 

###############################################################################
# NAME:        wait_mrp_active
#
# DESCRIPTION: 
#   Wait for $MRP_START_TIMEOUT minutes to confirm that the MRP is active.  If
#   we can't detect an active MRP, abort the script with an error.
#
# INPUT(S):
#   Arguments:
#     $1: database user
#     $2: user password
#     $3: tns service name
#     $4: database unique name
#
#   Globals:
#     None
#
# RETURN:
#   None
#
###############################################################################
wait_mrp_active()
{
  display "confirming media recovery is running"
  l_wma_status=1
  l_wma_curtime=`perl -e 'print int(time)'`
  l_wma_exptime=`expr $MRP_START_TIMEOUT "*" 60`
  l_wma_maxtime=`expr $l_wma_curtime + $l_wma_exptime`
  while [ "$l_wma_curtime" -lt "$l_wma_maxtime" ]
  do
    is_mrp_running $1 $2 $3 $4
    if [ "$?" -gt "0" ]; then
      l_wma_status=0
      break
    fi

    sleep $MRP_START_INTERVAL
    l_wma_curtime=`perl -e 'print int(time)'`
  done
  chkerr $l_wma_status "could not detect an active MRP after $MRP_START_TIMEOUT minutes"
}

split variable (eg. filename) into separate variables using set

FILENAME="split_this_into_bits"
set $(echo ${FILENAME} | sed 's/_/ /g')
echo $4 $3 $2 $1
bits into this split

Centre align text on a line in bash shell

Doesn't work in Korn shell due to %*s 

#COLUMNS=$(tput cols)   # width of the current window
COLUMNS=80
title="Hello world!"
printf "%*s\n" $(((${#title}+$COLUMNS)/2)) "$title"

and as a one-liner 

printf "%*s\n" $(( ( $(echo $* | wc -c ) + 80 ) / 2 )) "$*"

This works in Korn shell... 

TITLE="$1"
LINEWIDTH=80

LPAD=$(((${#TITLE}+$LINEWIDTH)/2))
printf %${LPAD}s "$TITLE"

Right justify text (or repeat characters) on a line with leader dots!

function rpadwait {
    text=$1
# -------------------------
# suppress newline for echo
# -------------------------
N=
C=
if echo "\c" | grep c >/dev/null 2>&1; then
    N='-n'
else
    C='\c'
fi


    echo ${N} "${text}${C}" | sed -e :a -e 's/^.\{1,80\}$/&\./;ta'
}

cron jobs not submitted - daemon run out of child processes?

Sometimes have problems with NFS mounts and this causes cron jobs to hang. If they are scheduled to run regularly, eventually cron will no longer be able to start any more jobs.

  • Check the cron log /var/adm/cron/log to see if there are any errors or other messages around the time the jobs should run.

If cron has hit its process limit (default 100), it will try again after a number of seconds (default 60).
Both the number of jobs and wait time are configured in the file /var/adm/cron/queuedefs. If it is unusual for cron to be running so many jobs, you can check the process table to view the jobs cron has created. These jobs will have parent process id (PPID) of the cron daemon.

$ ps -ef | grep cron | grep -v grep
    root  6750314        1   0   Apr 24      -  3:39 /usr/sbin/cron
solax025:root[/home/root]# ps -T 6750314
      PID    TTY  TIME CMD
  6750314      -  3:39 cron
 21168296      -  0:00     \--bsh
 58982414      -  0:00         \--sadc

In this example, we only have 1 job running

Find long-running processes with a cron job

Processes running longer than 24 hours have a date instead of a start time... 

58 08,14 * * * /home/ibmtools/scripts/oracle/dosh -vc "ps -ef|grep 'ibmtools/scripts/oracle'|perl -nae 'print if \$F[4] !~/:/'" >/tmp/lrp.txt; [[ $(grep -c ibmtools /tmp/lrp.txt) -ne 0 ]] && cat /tmp/lrp.txt|mailx -s '*** long running processes - please check ***' [email protected]

Process command line arguments in shell

Borrowed from Mozilla Firefox installer 

## Command line arg defaults
##
moz_debug=0
moz_debugger=""
moz_debugger_args=""
#
##
## Parse the command line
##
while [ $# -gt 0 ]
do
  case $1 in
    -g | --debug)
      moz_debug=1
      shift
      ;;
    -d | --debugger)
      moz_debugger=$2;
      if [ "${moz_debugger}" != "" ]; then
	shift 2
      else
        echo "-d requires an argument"
        exit 1
      fi
      ;;
    -a | --debugger-args)
      moz_debugger_args=$2;
      if [ "${moz_debugger_args}" != "" ]; then
	shift 2
      else
        echo "-a requires an argument"
        exit 1
      fi
      ;;
    *)
      break;
      ;;
  esac
done
#
##
## Program name given in $1
##
if [ $# -gt 0 ]
then
	MOZ_PROGRAM=$1
	shift
fi
##
## Program not given, try to guess a default
##
if [ -z "$MOZ_PROGRAM" ]
then
	##
	## Try this script's name with '-bin' appended
	##
	if [ -x "$MOZ_DEFAULT_NAME" ]
	then
		MOZ_PROGRAM=$MOZ_DEFAULT_NAME
	##
	## Try mozilla-bin
	## 
	elif [ -x "$MOZ_APPRUNNER_NAME" ]
	then
		MOZ_PROGRAM=$MOZ_APPRUNNER_NAME
	fi
fi
#
#
##
## Make sure the program is executable
##
if [ ! -x "$MOZ_PROGRAM" ]
then
	moz_bail "Cannot execute $MOZ_PROGRAM."
fi
#

Carry XWindows settings across sessions

# ----------------------
# push XWindows settings
# ----------------------
[[ "$(uname)" == "SunOS" ]] && PATH=/usr/openwin/bin:$PATH
WHOAMI=$(id | awk -F')' '{print $1}' | awk -F'(' '{print $2}')
xauth list > /tmp/xauth_list_$WHOAMI
chmod 777 /tmp/xauth_list_$WHOAMI

echo $DISPLAY > /tmp/xdisplay_$WHOAMI
chmod 777 /tmp/xdisplay_$WHOAMI

Cross-platform version of whoami

WHOAMI=$(id | awk -F')' '{print $1}' | awk -F'(' '{print $2}')

Set terminal title from command line

Put something like this in the .profile
-n do not output the trailing newline
-e enable interpretation of backslash escapes
|0 sets title of window and icon
|1 sets title of icon only
|2 sets title of window only

echo -en "\033]0;`hostname`\007"

echo -en "\033]2;`hostname`\007"

Remove blank lines and comments (also indented ones) from a file

awk -F\: '!/^($|[:space:]*#)/ {print $2}' /etc/oratab | sort | uniq

or as part of a script that removes comments and blank lines from all Korn shell scripts in a directory 

#!/usr/bin/ksh
for i in $(ls *ksh); do
    perl -p -i -e 's/^\s*#[^!]*$//; s/^\s*$//' $i
done

Return elements of an array in Korn shell

From here
Could be used to separate the columns of an SQL select when returning to the shell
This approach eliminates the need to put quotes around text with spaces in it. 

echo $KSH_VERSION
x="Red,Yellow is a color,Blue"
oIFS=$IFS
IFS=,
y=($x)
IFS=$oIFS
echo ${y[1]}

A decent Unix Prompt

export PS1="`uname -n`:`whoami`[\${PWD}]# "
or
export PS1='($?) $'ORACLE_SID" "`whoami`"@"`uname -n`":"'$'PWD"> "
export EDITOR=vi

Simple arithmetic

pipe the calculation into the shell calculator 

space_in_kb=$(echo $1 / 1024 | bc)

Calculate the remainder (modulo) of a division calculation 

if [[ $(echo "${NUMBER} % 2" | bc) -eq 0 ]]; then
    echo "${NUMBER} is even"
else
    echo "${NUMBER} is odd"
fi

or do it in awk if scientific notation maybe involved 

function calc { awk "BEGIN{print $*}"; }

if [[ $(calc "${SPACE_USED} + ${SPACE_AVAILABLE} - ${DATABASE_SIZE") -le 0 ]]; then
    echo "NOK"
fi

Script encryption and passphrase protection

Encrypt a shell script with the ability to execute the encrypted version 

scrypt(){ [ -n "$1" ]&&{ echo '. <(echo "$(tail -n+2 $0|base64 -d|mcrypt -dq)"); exit;'>$1.scrypt;cat $1|mcrypt|base64 >>$1.scrypt;chmod +x $1.scrypt;};}

Virtual host configuration in Apache http.conf

<VirtualHost *:80>
        ServerName dbamon
        DocumentRoot "/Volumes/data/Sites/dbamon_push"
        <Directory "/Volumes/data/Sites/dbamon_push">
                Options Includes FollowSymLinks
                AllowOverride All
                Order allow,deny
                Allow from all
        </Directory>
</VirtualHost>

and in /etc/hosts, add... 

127.0.0.1  dbamon

Mount a website (or any other remote resource) locally using WebDav

Redhat/CentOS 

yum install fuse-davfs2
or
yum install wdfs.x86_64

Debian 

apt-get install davfs2

then... 

sudo mkdir /mnt/webdav  # or whatever you'd like to call the directory
sudo mount.davfs [-o option[,option]...] device mount_point

In man's terms, that last line would translate to: 

id  #suppose this returns uid=501 and gid=502
sudo mount.davfs -o 501,502 https://your/web/site /mnt/webdav

Mac OSX 

osascript -e ' mount volume "http://username:[email protected]:portnum" '
or
osascript -e ' mount volume "http://[email protected]:portnum" '
or
osascript -e ' try mount volume "http://webdav.address:portnum" '
or
mount -t webdav http://webdav.address:portnum /mnt/webdav  # this one won't show up in the Finder Sidebar.

Add current hostname to list of hosts on an xcat server

CMD="nodels"
HOST=`hostname`
(echo "$HOST"; $CMD) | while read server
do
    echo "server:$server"
done

What is todays 'Day Of the Year' number?

DOY=`perl -e 'print sub{$_[7]}->(localtime)+1;'`

Convert Julian day numbers to dates

for day in 8 33 36 61 64 91 96 121 126 152 155 182 187 215 218 244 247 274 279 306 309 335 338 365; do date -d "`date +%Y`-01-01 +$(( ${day} - 1 ))days" +%d-%m-%Y; done

Send crontab job output to a date formatted log file

This will run a job every 5 minutes and send the output to a file ending with a time in hours and minutes.
The thing to note is the escaped percent signs. This is because a % sign is interpreted by cron to mean a newline character. Everything after the first % is treated as input to the program. 

*/5 *  * * * /var/www/cgi-bin/dbamon_collector.ksh       >/tmp/dbamon_collector.log.$(date "+\%H\%M") 2>&1

Edit crontab file without crontab -e

It can happen that you need to add or modify a line in the crontab of many users or across many servers at once.
In principle, there's nothing wrong with modifying the crontab file directly. You just lose the advantages of file locking (and syntax checking) that crontab -e offers.
Here we take a backup of the current crontab, print it out, echo an extra command and ask cron to use these as input (thus overwriting the existing crontab file). Just don't run this close to midnight :-) 

crontab -l > /tmp/crontab.`date '+%Y%m%d'`
(
cat /tmp/crontab.`date +'%Y%m%d'`
echo "02 10 * * * /home/ibmtools/scripts/oracle/export_parfile.ksh -s SID -f JDBEOP1.parfile"
) | crontab -

or 

crontab -l > /tmp/crontab.backup
crontab -l > /tmp/crontab.$$
perl -p -i -e 's!backup_send_tsm_dump!backup_export2tsm!g' /tmp/crontab.$$
crontab /tmp/crontab.$$
rm /tmp/crontab.$$

or 

crontab -l >$HOME/crontab.$(date '+%Y%m%d')
crontab -l | perl -p -e 's|/nas/software/oracle/scripts|/oracle/scripts|' | crontab

Use shell to convert a number in scientific notation to normal

var2convert='1.2345678e3'
printf -v var2convert "%.f" $var2convert
echo $var2convert  # magic!

Check for jobs running longer that 24 hours

Run from the management server across all Unix servers. Checks the 5th column in a ps listing. If it doesn't find a : (time separator), the process is running longer than 24 hours. 

# ---------------------------------------
# report on long running oraibm processes
# ---------------------------------------
01 17,14 * * * rm -f /tmp/lrp.txt;/home/ibmtools/scripts/oracle/dosh -vc \"ps -ef|egrep 'oraibm|scripts/oracle'>/tmp/lrp.txt;perl -nae 'print if \$F[4] !~/:/' /tmp/lrp.txt\" >>/tmp/lrp.txt;[[ $(egrep -c 'oraibm|scripts/or
acle' /tmp/lrp.txt) -ne 0 ]] && cat /tmp/lrp.txt|mailx -s '*** long running processes - please check ***' [email protected]

Global find and replace with perl (restricted to list of files provided by Unix find command)

find . -type f -exec perl -i -pe 's/something/else/g' {} \;

perl function to make filenames lower case

function lower { 
   perl -e 'for (@ARGV) { rename $_, lc($_) unless -e lc($_); }' * 
}

From the management server, search the TNS listener port for each database on a server and make an inline substitution in the ITM config files!

for i in `/home/ibmtools/scripts/oracle/dosh -c "ls -al /opt/IBM/ITM/config/*rz*cfg|grep -v lrwx"|awk '{print $NF}'`; do
    server=`echo $i|cut -d_ -f1 | awk -F'/' '{print $NF}'`
    db=`echo $i|cut -d'.' -f1 | awk -F'_' '{print $NF}'`
    OH=`ssh $server grep "^$db" /etc/oratab|cut -d: -f2`
    LISTENERPORT=`ssh $server cat $OH/network/admin/listener.ora|perl -00 -ne 'print $1 if /'$db'.*PORT.*=.*(\d{4})/s'`
    ssh $server perl -p -i.bak -e 's/1521/'$LISTENERPORT'/' $i
    ssh $server ls -al ${i}*
done

Run a job from cron every Nth day of the month

Example. Execute a job every third Saturday of the month.
Paste this into a file called calenday and put it in /usr/local/bin so it's (probably) on the PATH 

#!/usr/bin/ksh

# ix is the week number of the month ("2"nd Friday of the month, "3"rd Tuesday of the month)
# dy is the day number in Unix format (0 for Sunday, 1 for Monday, ... 6 for Saturday)
# eg. "calenday 3 6" returns the date of 3rd Saturday of the month.

ix=$1
dy=$2
SCHAR=$((($dy*2)+$dy+1))
ECHAR=$(($SCHAR+1))
cal `date '+%m %Y'` | egrep "\<[0-9]{1,2}\>" | cut -c${SCHAR}-${ECHAR} | xargs | awk {'print $'$ix'}'

Now in crontab, you should be able to do something like this:

15 20 * * * [[ `calenday 3 6` -eq `/bin/date '+%d'` ]] && su - oracle -c "run_my_backup.ksh"

This will also work on some Unices..
We send backups to a special TSM node on the second Friday of each month. This report must run a day later - but that is not guaranteed to be the second Saturday or even the 3rd. So... 

30 12 8-14 * 5 sleep 86400 && su - oracle -c "/usr/bin/perl -ne 'print if /ORX_M_SOL/ .. /^STOP/' /home/oracle/incoming/dbamon_spool_tsm_*.SOL | grep Archive | grep -v Client | mailx -s 'Monthly TSM backups' [email protected]"

ps listing does not show start time after 24 hours

But you can see elapsed time using your own ps command 

/usr/bin/ps -eo etime,user,pid,ppid,cpu,start,tty,time,args|tail -n +2|sort

ps -ef cuts off args cmd column on Solaris

To see more than 80 characters of the last column on Solaris

This shows all the individual arguments to the command 

pargs <pid>

This shows the ps listing in 'compatibility' mode (there are more compatibility commands in /usr/xpg4/bin) 

/usr/ucb/ps auww

To see more than 80 characters of the last column on AIX

This shows the full argument listing of a process (NOTE: no minus sign!) 

ps eww <pid>

Remove blank / empty lines from vi

Maybe you cut and pasted a file from Windows and it's full of blank lines and Control-M's now
There are several methods but I think this is the easiest to remember 

:g/^$/d

Right pad a variable

a function like rpad in SQL but for Shell

function rpad {
text=$1
padwidth=$2
padchar=$3
echo "$text" | sed -e :a -e 's/^.\{1,'$padwidth'\}$/&\'$padchar'/;ta'
}

Connect to a Windows server from Linux using rdesktop

My remmina stopped working so rolled my own. Very simple really. Put this is a shell. 

tsocks rdesktop -z -P -x m -a 16 -d MAIND -u sbarkley -p ****** -r disk:SOL=$HOME/Documents/SOL -g 95% 150.251.112.38 &

where...
-z    - enables compression
-P    - enables bitmap caching (saves network traffic)
-x m  - disables eye-candy features
-a 16 - reduce colour pallete to 16 colours
-d    - domain to connect to
-u    - username
-p    - password
-r    - setup a shared folder
-g    - geometry (use W x H or percentage)

Slight problems with rdesktop not working 100% of the time. Now using xfreerdp. Seems better... 

xfreerdp -g 90% --ignore-certificate --gdi sw -K -d wdcrhbp05 -u oraibm -p "`cat $HOME/scripts/.oraibm.password`" -T "wdcrhbp05_oraibm" --plugin cliprdr --plugin rdpdr --data disk:SOL:/home/bey9at77/Documents/SOL -- --plugin rdpsnd --data alsa -- 150.251.112.25 &

Reset your password bypassing password rules

must be done as root 

echo "user:new_password" | chpasswd

Sum the sizes of all files of an ls listing

It'll check to see if the sum of filesizes corresponds with the df -g (or h) listing (can get messed up due to open but deleted files) 

cd /oracle/export
df -g .
find . -name "*dmp*" -ls |  awk '{ SUM += $7} END { print SUM/1024/1024/1024 }'

Mount an iso image under Linux

mkdir -p /mnt/cdrom
mount -o loop /dev/cdrom /mnt/cdrom

How many processors on the machine?

  • AIX
lsdev -C|grep Process|wc -l
  • Solaris
psrinfo -v|grep "Status of processor"|wc -l
  • Linux
cat /proc/cpuinfo|grep processor|wc -l

Quick, simple, understandable example of how to use RRDTool

Use expect to respond automatically to interactive programs

#!/usr/bin/expect -f
spawn /usr/tivoli/tsm/client/oracle/bin64/tdpoconf passw -tdpo_optfile=/oracle/[lindex $argv 0]/admin/tdpo.opt
expect "assword:" {send "password\r"}
expect "assword:" {send "password\r"}
expect "assword:" {send "password\r"}

Use expect to allow file copy with scp (if ssh keys are not an option)

#!/usr/bin/expect -f
spawn scp /home/oracle/.profile oracle@hn512:/tmp/prfl
set pass "thepassword"
expect {
        password: {send "$pass\r" ; exp_continue}
        eof exit
}

Within a shell script set up simultaneous output to both terminal and a file using a FIFO (named pipes)

or "How to send shell output to screen/stdout as well as to a logfile using tee and redirection with exec"

examples here on unix.com 

#!/bin/ksh

REDIR=test.redir
FIFO=test.pipe
[[ -e $FIFO ]] || mkfifo $FIFO

# make a new channel(3) and copy channel 1's destination as its own (does NOT POINT TO channel 1's destination)
# this allows the normal output to continue to STDOUT but also get printed to whatever file is attached to channel 3
exec 3>&1

# anything coming in on the pipe, send it to $REDIR and to channel 3
tee $REDIR <$FIFO >&3 &

# redirect STDOUT to the pipe
exec > $FIFO

echo "going to default output"
echo "forcing to channel 1" >&1
echo "forcing to channel 2" >&2
echo "forcing to channel 3" >&3

More elaborate example

found here on stackoverflow 

#!/bin/sh

# Author: Harvey Chapman <hchapman _AT_ 3gfp.com>
# Description: POSIX shell functions that can be used with tee to simultaneously put
#              stderr and stdout to both a file and stdout
#
# Based on:
#    Re: How to redirect stderr and stdout to a file plus display at the same time
#    http://www.travishartwell.net/blog/2006/08/19_2220

#
# Original example function from Travis Hartwell's blog.
# Note: I've made minor changes to it.
example()
{
  OUTPUT_LOG=output.log
  OUTPUT_PIPE=output.pipe

  # This should really be -p to test that it's a pipe.
  if [ ! -e $OUTPUT_PIPE ]; then
      mkfifo $OUTPUT_PIPE
  fi

  # This should really be -f to test that it's a regular file.
  if [ -e $OUTPUT_LOG ]; then
      rm $OUTPUT_LOG
  fi

  exec 3>&1 4>&2
  tee $OUTPUT_LOG < $OUTPUT_PIPE >&3 &
  tpid=$!
  exec > $OUTPUT_PIPE 2>&1

  echo "This is on standard out"
  echo "This is on standard err" >&2

  exec 1>&3 3>&- 2>&4 4>&-
  wait $tpid

  rm $OUTPUT_PIPE
}

# A slightly reduced version of example()
example2()
{
  OUTPUT_LOG=output.log
  OUTPUT_PIPE=output.pipe

  rm -f $OUTPUT_PIPE
  mkfifo $OUTPUT_PIPE
  rm -f $OUTPUT_LOG

  tee $OUTPUT_LOG < $OUTPUT_PIPE &
  tpid=$!

  exec 3>&1 4>&2 >$OUTPUT_PIPE 2>&1

  echo "This is on standard out"
  echo "This is on standard err" >&2

  exec 1>&3 3>&- 2>&4 4>&-
  wait $tpid
  rm -f $OUTPUT_PIPE
}

#
# Logging methods based on above. See the example below for how to use them.
#

# Usage: start_logging [delete_existing_logfile]
start_logging()
{
  # Check to see if OUTPUT_LOG and OUTPUT_PIPE need to be defined.
  if [ -z "$OUTPUT_LOG" ]; then
    OUTPUT_LOG=output.log
  fi
  if [ -z "$OUTPUT_PIPE" ]; then
    OUTPUT_PIPE=output.pipe
  fi
  # Make sure that we're not already logging.
  if [ -n "$OUTPUT_PID" ]; then
    echo "Logging already started!"
    return 1
  fi

  # Always remove the log and pipe first.
  rm -f $OUTPUT_PIPE
  # Delete the logfile first if told to.
  if [ "$1" = delete_existing_logfile ]; then
    rm -f $OUTPUT_LOG
  fi

  mkfifo $OUTPUT_PIPE
  tee -a $OUTPUT_LOG < $OUTPUT_PIPE &
  OUTPUT_PID=$!

  exec 3>&1 4>&2 >$OUTPUT_PIPE 2>&1
}

stop_logging()
{
  # Make sure that we're currently logging.
  if [ -z "$OUTPUT_PID" ]; then
    echo "Logging not yet started!"
    return 1
  fi
  exec 1>&3 3>&- 2>&4 4>&-
  wait $OUTPUT_PID
  rm -f $OUTPUT_PIPE
  unset OUTPUT_PID
}

example3()
{
  start_logging
  #start_logging delete_existing_logfile
  echo "This is on standard out"
  echo "This is on standard err" >&2
  stop_logging
}

RedHat root filesystem has gone read-only

Kill all processes for a user

for prc in `ps -ef | grep -E "^ +[o]racle" | awk '{print $2}'`; do
    kill $prc
done

isRGHere

Checks if resource group is on this leg of an HACMP cluster. Returns 0 if true else 1. 

#!/usr/bin/ksh

SCRIPT=`basename $0`

function rctest
{
exitcode=$1
msg=$2
if [ $exitcode -ne 0 ]; then
        echo "********************************************************************************"
        echo "\nScript $SCRIPT finished with errors."
        echo "$msg."
        echo "Returncode : $exitcode."
        echo "\n********************************************************************************"

        fi
exit $exitcode

}

RGINFO=/usr/es/sbin/cluster/utilities/clRGinfo
[[ ! -f $RGINFO ]] &&  rctest 1 "clRGinfo not found"

if [ $# -eq 1 ]
then
        RG=`echo $1 | cut -c 1-14`
else
        rctest 10  "Usage: `basename $0` <RG name>"
fi

$RGINFO |grep -qwp $RG ||  rctest 9 "$RG is not defined"

THISNODE=`/usr/es/sbin/cluster/utilities/get_local_nodename | cut -c 1-14`
$RGINFO |grep -wp $RG |grep -w $THISNODE |grep -wq ONLINE

AIX: add a user to a group

chgrpmem -m + oracle ibmtls

Find swear words in scripts by comparing it to a list found on the internet

wget "http://www.bannedwordlist.com/lists/swearWords.csv" -O /tmp/s ; for word in $(cat /tmp/s | sed -e 's/ /_/g' -e 's/,/ /g') ; do grep -wR $word *; done | less

grep -p (for paragraph) works on AIX but not on Linux or Solaris

Use awk instead 

awk 'BEGIN {FS="\n" RS="\n\n"} /search pattern/ { do something }' <file>
/usr/xpg4/bin/awk 'BEGIN {RS="\n\n";FS="\n"} /AGRHDWQ1/ {print $2}' dsm.sys | awk '{print $NF}'

this prints the last word of the second line in the paragraph in dsm.sys that contains the search term AGRHDWQ1.
Or slightly simpler... 

awk -v RS='' '/NHAPPLP1/' /etc/tsm/dsm.sys     # (use /usr/xpg4/bin/awk on Solaris)

or, case insensitively: 

awk -v RS='' 'tolower($0) ~/so_u_clubqa_orx_d_cab/' /etc/tsm/dsm.sys

Using -v means you don't have to use a BEGIN section.

debug/redirect log of a shell

#!/usr/bin/ksh
exec 2>/tmp/mytest

Different ways of Iteration in korn shell with a while loop

IFS="|"
exec 0<$statfile
while read host db started stopped
do
    rrdfile="export_duration_${host}_${db}.rrd"
    $RRDTOOL update ${RRDDIR}/${rrdfile} $started:$started $stopped
done

or 

while read host db started stopped
do
    rrdfile="export_duration_${host}_${db}.rrd"
    $RRDTOOL update ${RRDDIR}/${rrdfile} $started:$started $stopped
done <statfile

or 

    cat $statfile | sort -nk4 | while IFS=\| read host db type started stopped
    do
        [[ "$stopped" == "" ]] && continue
        rrdfile="export_duration_${host}_${db}.rrd"
        $RRDTOOL update ${RRDDIR}/${rrdfile} ${started}:${started}:${stopped}
        [[ $? -ne 0 ]] && echo "nok: $?"
    done

Filesystem 100% full, what's taking up all the space?

find /oracle/endur -xdev -ls|sort -nr +6|head

or 

/dev/esb01fs010001      1.00      0.00  100%     1838    69% /oracle
beuxdsysesb01:root[/root]# cd /oracle
beuxdsysesb01:root[/oracle]# du -gsx * | sort -n
0.00    checkpoints
0.00    flash_recovery_area
0.00    lost+found
0.00    oraInst.loc
0.00    oraInventory
0.09    admin
0.99    diag

cd diag and repeat until culprits are found

Show paged memory hogs on AIX

svmon -Pt20 | perl -e 'while(<>){print if($.==2||$&&&!$s++);$.=0 if(/^-+$/)}'

Capitalise the first letter of a word in Korn Shell

Not as easy as it sounds if it needs to work in non-ksh93 shells 

function capit { 
    typeset -u first
    first=${1%"${1#?}"}
    printf "%s\n" "${first}${1#?}"
    return 0
}

or maybe more neatly done in Perl...
This will capitalise each word in a sentence passed into it 

function capit { 
    echo "$1" |  perl -pe "s/([\w']+)/\u\L\1/g"
    return 0
}

Remove / rename a directory / file containing weird control characters

Use ls with -i to see inode listing 

ls -bali

Use find with -inum to get the filename and -exec to remove it 

find . -inum <inode from ls listing> -exec rm -f {} \;

Run a local script on a remote host

Doing it this way means we do not need to scp the shell over to the host before executing it! 

ssh user@host 'sh' < your_script.sh

First of all, this command is a redirection: your shell will open the file your_script.sh and feed it as input to the ssh command. ssh, in turn, will tunnel its stdin to the remote command, namely, sh instance. And sh without arguments reads its script from stdin. So we got sh instance, which is launched on remote host, but reads commands from your local file.

Get a list of running instances - but only those started by the current user

This is the way I wanted to do it but there's an issue. Where does that spare } character come from? Anyone? 

ps -ef | grep [p]mon | awk -v dbowner=$(id -un) '{ if ($1==dbowner) { gsub(/ora_pmon_/,"",$NF); print $NF; } }'

ps listing 

$ ps -ef | grep [p]mon
  oracle 13304024        1   0   Jun 07      -  2:39 ora_pmon_reportk
  oracle 26018178        1   0   Jun 07      -  3:01 ora_pmon_dwh_perf
  oracle 29229432        1   0   Jun 07      -  2:30 ora_pmon_adso
 oraebso 18022994        1   0   Jun 07      -  2:38 ora_pmon_EBSO
  oracle 30278192        1   0   Jun 07      -  2:48 ora_pmon_owb112k

Results of above command 

reportk
dwh_perf
adso
}
owb112k

Workaround 1. Send the ps listing to a file and work on it without pipes. Works but it's a bit long-winded especially as we have to clean up afterwards. 

ps -ef | grep [p]mon>/tmp/results.$$ 2>/dev/null; awk -v dbowner=$(id -un) '{ if ($1==dbowner) { gsub(/ora_pmon_/,"",$NF); print $NF; } }' /tmp/results.$$; rm -f /tmp/results.$$

Workaround 2. Don't like the grep -v but until I find out where that } is coming from.. 

ps -ef | grep [p]mon | awk -v dbowner=$(id -un) '{ if ($1==dbowner) { gsub(/ora_pmon_/,"",$NF); print $NF; } }' | egrep -v 'grep|}|ASM'

Return a list of Oracle databases running on a list of remote servers

#!/bin/ksh

serverlist=`cat /home/tools/ini/system/oracle_servers | sort -n | tr "\n" " "`

if [ -z "${serverlist}" ]
then
    echo "no servers found" && exit 1
fi


for server in ${serverlist}
do
    ssh ${server} "ps -ef | grep [o]ra_pmon_" >/tmp/${server}.dblist
done

for server in ${serverlist}
do
    cat /tmp/${server}.dblist | awk -F_ -v SRV=${server} 'BEGIN {print SRV ":"} {print $NF}' | tr "\n" " "
    echo
done

Clever trick to check whether SQL output needs sending to someone

Using an exit status from SQL*Plus back to the shell so we can decide what to do. Nice one! 

#!/bin/bash
tempfile=/tmp/audit_locked_accounts_$ORACLE_SID.txt

# Start sqlplus and check for locked accounts
sqlplus -S "/ as sysdba" << EOF > $tempfile
set pages 0
select 'The following accounts were found to be unlocked and should not be'
from dual;

def exit_status = 0
col xs new_value exit_status

select username
,      account_status
,      1 as xs
from   dba_users
where  account_status != 'LOCKED'
and    username in ('HR', 'SCOTT', 'OUTLN', 'MDSYS', 'CTXSYS')
/
exit &exit_status
EOF

# If the exit status of sqlplus was not 0 then we will send an email
if [ $? != 0 ]; then
    mail -s "Accounts Unlocked in $ORACLE_SID" oracle < $tempfile
fi

Check RMAN logfiles for errors from cron every day

00 09 * * * /home/tools/scripts/oracle/dosh 'find /home/tools/logs/rman -name "*online.log" -mtime -1 -exec sed -ne "/^RMAN-/,/^$/p" {} \\; -ls' | mailx
 -s 'RMAN errors last night' [email protected]

/home/ibmtools/scripts/oracle/dosh 'find /oracle/export -name "expdp*log" -mtime -1 -exec grep ORA- {} \; -ls' | mailx -s 'Datapump errors for Baxter last night' [email protected]

An alternative to getopts

These methods have the advantage of allowing arguments of any length not just 1 character 

countArgs=$#
while [[ $# -gt 0 ]]
do
    idx1=$($ECHO $1|$GREP "="|wc -l)
    if [[ $idx1 -gt 0 ]]; then
        key=$($ECHO $1|$CUT -d '=' -f 1)
        value=$($ECHO $1|$CUT -d '=' -f 2)
    else
        key=$($ECHO $1|$CUT -d '=' -f 1)
        idx2=$($ECHO $1|$GREP "^-"|wc -l)
        if [[ $idx2 -eq 0 ]]; then
            $ECHO -e "\n ERROR: $1 is an unsupported argument passed to agentDeploy.sh.\n"  
            usage
            exit 1
        fi
    fi

    case "$key" in
    -help)
        if [[ $upgradeFlag ]]; then
            upgradeUsage
        else
            freshUsage
        fi
        exit 0;;
    AGENT_BASE_DIR)
        agentBaseDir=$($ECHO $value|$SED 's/\/$//')
        checkBaseFlag=TRUE;;
    OMS_HOST)
        omsHost=$value
        checkOMSHost=TRUE;;
    EM_UPLOAD_PORT)
        omsPort=$value
        checkOMSPort=TRUE;;
    AGENT_INSTANCE_HOME)
        instHome=$($ECHO $value | $SED 's/\/$//');;
    AGENT_REGISTRATION_PASSWORD)
        pswd=$value;;
    s_encrSecurePwd)
        pswd=$value;;  
    RESPONSE_FILE)
        rspFlag=TRUE  
        rspLoc=$value;;
    OLD_AGENT_VERSION)
        oldAgentVersion=$value;;            
    OLDHOME)
        oldHome=$value;;    
    -debug)
        debugSwitch=TRUE;;
    -forceConfigure)
        forceFlag=TRUE;;
    -configOnly)
        configFlag=TRUE    
        validationFlag=TRUE;;
    -agentImageLocation)
        archiveLoc=$value
        checkArchiveFlag=TRUE;;
    -invPtrLoc) shift
        ptrLoc=$1;;
    -Upgrade)
        UpgradeFlag=TRUE
        validFlag=TRUE;;
    INVENTORY_LOCATION)
        validFlag=TRUE;;
    b_forceInstCheck)
        validFlag=TRUE
        forcefullFlag=TRUE;;
    -prereqOnly)
        validationFlag=TRUE
        prereqFlag=TRUE;;
    -executeRootsh)
        validFlag=TRUE;;
    *)  idx=$($ECHO $1|$GREP "^-"|wc -l)
        if [[ $idx -eq 0 ]]; then
            validFlag=TRUE
        else
            $ECHO -e "\n ERROR: Invalid argument $key passed."
            usage
            exit 1          
        fi
    esac
    shift
done

another way - found in /etc/init.d/functions (daemon function) in Fedora 

    while [ "$1" != "${1##[-+]}" ]; do
        case $1 in
        '')
            echo $"$0: Usage: daemon [+/-nicelevel] {program}" "[arg1]..."
            return 1
            ;;
        --check)
            base=$2
            gotbase="yes"
            shift 2
            ;;
        --check=?*)
            base=${1#--check=}
            gotbase="yes"
            shift
            ;;
        --user)
            user=$2
            shift 2
            ;;
        --user=?*)
            user=${1#--user=}
            shift
            ;;
        --pidfile)
            pid_file=$2
            shift 2
            ;;
        --pidfile=?*)
            pid_file=${1#--pidfile=}
            shift
            ;;
        --force)
            force="force"
            shift
            ;;
        [-+][0-9]*)
            nice="nice -n $1"
            shift
            ;;
        *)
            echo $"$0: Usage: daemon [+/-nicelevel] {program}" "[arg1]..."
            return 1
            ;;
      esac
    done

another way - found in adstrtal.sh (middleware start script for EBS) 

#
# Parse Arguments
#

for nxtarg in $*
do
  arg=`echo $nxtarg | sed 's/^//'`

  case $arg in
    -secureapps)    if test "$secureapps" = "" ; then
                       secureapps=$arg
                    else
                       echo "$0: Duplicate Argument passed : $arg"
                       usage
                    fi
                ;;
    -nodbchk)       if test "$nodbchk" = "" ; then
                       nodbchk=$arg
                    else
                       echo "$0: Duplicate Argument passed : $arg"
                       usage
                    fi
                ;;
    -nopromptmsg)   if test "$nopromptmsg" = "" ; then
                       nopromptmsg=$arg
                    else
                       echo "$0: Duplicate Argument passed : $arg"
                       usage
                    fi
                ;;
    *)              if test "$unpw" = "" ; then
                       unpw=$arg
                    else
                       echo "$0: Duplicate Argument passed : $arg"
                       usage
                    fi
  esac
done

and another in adautocfg.sh 

for myarg in $*
do

  arg=`echo $myarg | sed 's/^-//'`
  case $arg in
    appspass=*)
            appspass=`echo $arg | sed 's/appspass=//g'`
            shift
            ;;
    nocustom)
            myparams="$myparams $arg"
            shift
            ;;
    noversionchecks)
            myparams="$myparams $arg"
            shift
            ;;
    promptmsg=*)
            promptmsg=`echo $arg | sed 's/promptmsg=//g'`
            shift
            ;;
        *)  echo "$0: unrecognized action specified"
            exit 1
  esac
done

Run a script on multiple servers

#!/bin/ksh
# ==============================================================================
# Name         : dosh   (Distributed Oracle SHell)
# Description  : Runs a command on all Oracle servers
#
# Parameters   : h - displays help
#                v - verbose (default, like all unix commands is silent)
#                c - command to be executed
#
# Example      : ./dosh -v -c 'ls -al'
#
# Modification History
# ====================
# When      Who               What
# ========= ================= ==================================================
# 08-FEB-13 Stuart Barkley    Created
# ==============================================================================

# --------------------------------------------------------------------------
# this list of servers is generated by the DBAMON tool so will be up to date
# --------------------------------------------------------------------------
serverfile=/home/ibmtools/etc/oracle/oracle_servers
[[ -z "${serverfile}" ]] && echo "Server list ${serverfile} not found" && exit 1

serverlist=`cat ${serverfile} | sort -n | tr "\n" " "`


# -------------------------
# get the arguments, if any
# -------------------------
while getopts "hvc:" OPT
do
    case "$OPT" in
    h) echo "\nUsage: $0 [-v] [-h] -c '<command_to_be_executed_remotely>'\n";
       exit;
       ;;
    v) VERBOSE="Y";
       ;;
    c) CMMND=$OPTARG;
       ;;
    *) echo "\nUsage: $0 [-v] [-h] -c '<command_to_be_executed_remotely>'\n";
       exit;
       ;;
    esac
done
shift $((OPTIND-1))
VERBOSE=${VERBOSE:-"N"}


# --------------------------------
# check we have required arguments
# --------------------------------
[[ -z $CMMND ]] && echo "Enter command to be executed like this: $0 -c 'ls -al'" && exit 1


# ----------------------------
# loop over the remote servers
# ----------------------------
for server in ${serverlist}
do
    if [[ "$VERBOSE" == "Y" ]]; then
        echo ${server} `date +'%d-%b-%Y %H:%M:%S'`
        echo "------------------------------------"
        set -x
    fi
    ssh ${server} "$CMMND"
    set +x
    [[ "$VERBOSE" == "Y" ]] && echo
done

to be investigated... this one might be cleverer than mine... 

tmpdir=${TMPDIR:-/tmp}/pssh.$$
count=0
while userhost; do
    ssh -n -o BatchMode=yes ${userhost} 'uname -a' > ${tmpdir}/${userhost} 2>&1 &
    count=`expr $count + 1`
done < userhost.lst
while [ $count -gt 0 ]; do
    wait $pids
    count=`expr $count - 1`
done
echo "Output for hosts are in $tmpdir"


improved version...
just escape double quotes and dollar signs... 

#!/bin/ksh
# ==============================================================================
# Name         : dosh   (Distributed Oracle SHell)
# Description  : Runs a command on all Oracle servers
#
# Parameters   : h - displays help
#                v - verbose (default, like all unix commands is silent)
#                c - command to be executed
#
# Example      : ./dosh -v -c 'ls -al'
#
# Modification History
# ====================
# When      Who               What
# ========= ================= ==================================================
# 08-FEB-13 Stuart Barkley    Created
# 08-JUL-15 Stuart Barkley    Background execution mode
# ==============================================================================
# --------------------------------------------------------------------------
# this list of servers is generated by the DBAMON tool so will be up to date
# --------------------------------------------------------------------------
serverfile=/home/ibmtools/etc/oracle/oracle_servers
[[ -z "${serverfile}" ]] && echo "Server list ${serverfile} not found" && exit 1

serverlist=`cat ${serverfile} | sort -n | tr "\n" " "`


# -------------------------
# get the arguments, if any
# -------------------------
while getopts "hbvc:" OPT
do
    case "$OPT" in
    h) echo "\nUsage: $0 [-b] [-v] [-h] -c '<command_to_be_executed_remotely>'\n";
       exit;
       ;;
    b) BACKGROUND="Y";
       ;;
    v) VERBOSE="Y";
       ;;
    c) CMMND=$OPTARG;
       ;;
    *) echo "\nUsage: $0 [-b] [-v] [-h] -c '<command_to_be_executed_remotely>'\n";
       exit;
       ;;
    esac
done
shift $((OPTIND-1))
VERBOSE=${VERBOSE:-"N"}


# --------------------------------
# check we have required arguments
# --------------------------------
[[ -z $CMMND ]] && echo "Enter command to be executed like this: $0 -c 'ls -al'" && exit 1


# ---------------------------------
# put the thing to be run in a file
# ---------------------------------
printf "%s\n" "$CMMND" > /tmp/dosh.$$


# ----------------------------
# loop over the remote servers
# ----------------------------
for server in ${serverlist}
do
    if [[ "$VERBOSE" == "Y" ]]; then
        printf "%s\n" ${server} `date +'%d-%b-%Y %H:%M:%S'`
        printf "%s\n" "------------------------------------"
        set -x
    fi
    scp -q /tmp/dosh.$$ ${server}:/tmp/
    if [[ "$BACKGROUND" == "Y" ]]; then
        ssh -n -o BatchMode=yes ${server} "ksh /tmp/dosh.$$; rm -f /tmp/dosh.$$"> /tmp/dosh_${server}.out 2>&1 &
    else
        ssh -n -o BatchMode=yes ${server} "ksh /tmp/dosh.$$; rm -f /tmp/dosh.$$"
    fi
    set +x
    [[ "$VERBOSE" == "Y" ]] && echo
done
rm -f /tmp/dosh.$$

[[ "$BACKGROUND" == "Y" ]] && printf "%s\n" "Type 'ls -altr /tmp/dosh*out' to see output from the commands"

How can a shell script find out what directory it is in?

basename $0 does not always give the desired answer (might give "./"! 

DIR="$(cd "$(dirname "$0")" && pwd)"

Array processing

array=(1 2 3)
unset array[2]
echo ${array[2]}          # null
indices=(${!array[@]})    # create an array of the indices of "array"
size=${#indices[@]}       # the size of "array" is the number of indices into it
size=${#array[@]}         # same
echo ${array[@]: -1}      # you can use slices to get array elements, -1 is the last one, etc.
for element in ${array[@]}; do    # iterate over the array without an index

for index in ${indices[@]}        # iterate over the array WITH an index
do
    echo "Index: ${index}, Element: ${array[index]}"
done

for index in ${!array[@]}         # iterate over the array WITH an index, directly

array+=("new element")    # append a new element without referring to an index
((counter++))             # shorter than ((counter=counter+1)) or ((counter+=1))
if [[ $var == 3 ]]        # you can use the more "natural" comparison operators inside double square brackets
while [[ $var < 11 ]]     # another example
echo ${array[${index}-1]  # math inside an array subscript

Send an email by talking directly to an smtp server

#!/bin/bash
telnet smtp.domain.com 25 <<EOTXT>>/tmp/smtp.log
HELO me.domain.com
MAIL FROM:<[email protected]>
RCPT TO:<[email protected]>
DATA
From: Stuart <[email protected]>
To: Anne <[email protected]>
Subject: testing smtp email

Hello, this should appear in the body
.
QUIT
EOTXT

Send an email by talking directly to an smtp server via file descriptor (no telnet! this time), adding authentication

#!/bin/bash

#
# mail.sh
#
# 2008 - Mike Golvach - [email protected]
# 2010 - Rayber
#
# Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License
#

if [ $# -ne 7 ]
then
echo "Usage: $0 FromAdress ToAdress Domain MailServer MailTextFile AuthEmail AuthPass"
exit 1
fi

from=$1
to=$2
domain=$3
mailserver=$4
mailtext=$5
authemail=`echo $6|openssl enc -base64|awk 'sub("..$", "")'`
authpass=`echo $7|openssl enc -base64|awk 'sub("..$", "")'`

if [ ! -f $mailtext ]
then
echo "Cannot find your mail text file. Exiting..."
exit 1
fi

exec 9<>/dev/tcp/$mailserver/25
echo "HELO $domain" >&9
read -r temp <&9
echo "$temp"
echo "auth login" >&9
read -r temp <&9
echo "$authemail" >&9
read -r temp <&9
echo "$authpass" >&9
read -r temp <&9
echo "Mail From: $from" >&9
read -r temp <&9
echo "$temp"
echo "Rcpt To: $to" >&9
read -r temp <&9
echo "$temp"
echo "Data" >&9
read -r temp <&9
echo "$temp"
cat $mailtext >&9
echo "." >&9
read -r temp <&9
echo "$temp"
echo "quit" >&9
read -r temp <&9
echo "$temp"
9>&-
9<&-
echo "All Done Sending Email. See above for errors"
exit 0

rsync examples

Also see Distribute files to multiple servers using rsync and ssh 

#!/bin/sh
ssh  <remote_host> '/bin/mkdir -p /etc /etc/rc.config.d /etc/security /etc/mail'
rsync --rsync-path /usr/bin/rsync -Liprogtz --out-format=%f%L  /etc/passwd /etc/passwd.post /etc/group /etc/hosts /etc/services /etc/resolv.conf /etc/exclude.rootvg <remote_host>:/etc
rsync --rsync-path /usr/bin/rsync -Liprogtz --out-format=%f%L  /etc/hosts.allow.xcat <remote_host>:/etc/hosts.allow
rsync --rsync-path /usr/bin/rsync -Liprogtz --out-format=%f%L  /etc/rc.config.d/sap <remote_host>:/etc/rc.config.d
rsync --rsync-path /usr/bin/rsync -Liprogtz --out-format=%f%L  /etc/security/group /etc/security/limits /etc/security/login.cfg /etc/security/passwd /etc/security/user <remote_host>:/etc/security
rsync --rsync-path /usr/bin/rsync -Liprogtz --out-format=%f%L  /etc/mail/sendmail.cf <remote_host>:/etc/mail

rsync -av --progress /home/ibmtools/scripts/oracle/* benouerp07:/home/ibmtools/scripts/oracle/

Handy aliases

Strip out comments and blank lines from a file 

alias strip='grep -Ev '\''^(#|$)'\'''

Does a ps and a grep 

alias psg='ps -ef | grep -v $$ | grep -i '

Getting a decent listing of filesystem space available. It is ordered such that the filesystems with no space left are at the end. 

OS=$(uname -s)
[[ "$OS" == "SunOS" ]] && alias dfg='df -h|sed -e '1d'|sort -n -k5|awk '\'' BEGIN {printf("%-35s%-10s%-10s%-6s%-30s\n","Filesystem","Total","Free","%Used","Mounted on")} {printf("%-35s%-10s%-10s%-6s%-30s\n",$1,$2,$4,$5,$6)}'\'''
[[ "$OS" == "AIX" ]]   && alias dfg='df -g|sed -e '1d'|sort -n -k4|awk '\'' BEGIN {printf("%-35s%-10s%-10s%-6s%-30s\n","Filesystem","Total GB","Free","%Used","Mounted on")} {printf("%-35s%-10s%-10s%-6s%-30s\n",$1,$2,$3,$4,$7)}'\'''

A handy way of listing subdirectories and their files 

alias filetree="find . -print | sed -e 's;[^/]*/;|____;g;s;____|; |;g'"

Watch progress of a copy 

alias cpProgress="rsync --progress -ravz"
Reboots Linksys router
alias rebootlinksys="curl -u 'admin:password' 'http://192.168.1.2/setup.cgi?todo=reboot'"

Nice one for bash. Colour codes the prompt depending on the outcome of the previous command 

bash_prompt_command()
{
    RTN=$?
    prevCmd=$(prevCmd $RTN)
}
PROMPT_COMMAND=bash_prompt_command
prevCmd()
{
    if [ $1 == 0 ] ; then
        echo $GREEN
    else
        echo $RED
    fi
}
if [ $(tput colors) -gt 0 ] ; then
    RED=$(tput setaf 1)
    GREEN=$(tput setaf 2)
    RST=$(tput op)
fi
export PS1="\[\e[36m\]\u.\h.\W\[\e[0m\]\[\$prevCmd\]>\[$RST\]"

Mmmm, to be looked into. Executes remote commands on a unix box using curl. 

#/bin/sh
#
# WAG320N-HACK
# Ver. 1.0
# 12/09/2010
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

# Set username and password in the form of "username:password"
# example: "admin:admin"
my_access="admin:admin"

# Parameters test
if [ -z "$1" ]
then
  echo "wag320n-hack.sh: missing remote command"
  echo "usage: wag320n-hack.sh <remote command>"
  echo "example: wag320n-hack.sh /bin/ls -la /usr/sbin"
  echo "Note: always use full path"
  echo ""
  echo "wag320n-hack.sh - Ver. 1.0 - 12/09/2010"
  echo "Licensed under GPL V. 3"
  echo ""
  exit 0
fi

# Get the command
my_command="ping_size="'$('"$@"' 1>&2)'

curl -s -G -u "$my_access" --data-urlencode 'todo=ping_test' --data-urlencode 'this_file=Diagnostics.htm' --data-urlencode 'next_file=Ping.htm' --data-urlencode 'c4_ping_ipaddr=192.168.1.1' --data-urlencode 'ping_timeout=5000' --data-urlencode 'ping_interval=1000' --data-urlencode 'ping_number=1' --data-urlencode "$my_command" http://192.168.1.1/setup.cgi | sed -e '/HTTP\/1.0 200 OK/q'

How to configure SSH with public/private keys

Use SSH config file

Host server10
  Hostname 1.2.3.4
  IdentityFile ~/.ssh/id_dsa
  user foobar
  Port 30000
  ForwardX11Trusted yes
  TCPKeepAlive yes

then just connect using 

ssh server10

A decent sed tutorial

From grymoire.com

A decent korn/bash shell tutorial

From dartmouth.edu Reproduced here just in case it disappears! Advanced shell scripting

trap

     Example Handling Traps With ksh - Discussion of the kill command


EXAMPLE TEMPLATE:


PRODUCT:    HP-UX 11iV1 Version B.11.11
            HP Tru64 V5.1B PK4
            Sun/Solaris SunOS V5.8
            Linux 2.6 kernel


COMPONENT:  ksh


SOURCE:     Philippe Vouters
            Fontainebleau/France


LOW COST HIGH-TECH PRODUCTS:  http://techno-star.fr


OVERVIEW:

The ksh script below shows how to eventually handle traps in the situation 
where someone might try to kill a script by killing individual commands run 
by that script or the entire process group a script is running in. The kill 
command (usually a shell builtin) may be used to send a signal to a process 
group (with the -<pid> syntax) or an individual process. The example ksh 
script below runs /bin/sleep as the foreground process, the example ksh 
scripts immediately returns when the /bin/sleep process has terminated. Most 
signals sent to the shell are ignored until after the foreground process 
terminates. This is in order to avoid creating zombie processes. Therefore a 
kill <pid> on the example ksh script waits for the termination of the 
/bin/sleep process.

The status value $? in the trap refers to the exit status of the command to run 
and therefore is the exit status of the /bin/sleep process. The called function 
in the trap handler shows how to correctly examine the effect of the kill 
command on the shell or it's children.

To examine the value of $? in a trap handler means that you must understand what
it can be set and how different signals delivered to either the shell or the 
foreground process (or the process group) might affect the value of $?.

The example shell script prints $? using echo but it does not perform tests on 
the value of $?. For a complete solution when attempting to trap signals in a 
shell you would also need code that examined the value of $? after the 
foreground process had completed.


*** CAUTION ***

This sample script has been tested using HP-UX B.11.11, HP Tru64 V5.1B PK4, 
SunOS V5.8 and Fedora Core 4 (homed version of Red Hat Linux).  However, we
cannot guarantee its effectiveness because of the possibility of error in 
transmitting or implementing it. It is meant to be used as a template for 
writing your own scripts, and may require modification for use on your system. 


SCRIPT NOTES:

To notice that the ksh script and /bin/sleep share the same process group
identifier (PGID), issue the following commands:

[philippe@victor ~]$ who
philippe :0           Jan 10 10:16
philippe pts/1        Jan 10 21:30 (:0.0)
philippe pts/2        Jan 10 21:30 (:0.0)
[philippe@victor ~]$ tty
/dev/pts/1
[philippe@victor ~]$ ps -j -t pts/2
  PID  PGID   SID TTY          TIME CMD
11072 11072 11072 pts/2    00:00:00 bash
11113 11113 11072 pts/2    00:00:00 ksh
11116 11113 11072 pts/2    00:00:00 sleep

In this case sending kill -INT -11113 will send SIGINT to the process group 
11113. Both of the ksh and sleep processes are contained within this process 
group.

Important Note:

On HP-UX, you have to $ export UNIX95=1 in order to be able to use the
-j option of the ps command.


SCRIPT:

                             COPYRIGHT (C) 2005 BY
                              HEWLETT-PACKARD COMPANY
                                ALL RIGHTS RESERVED.

     THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY BE USED AND COPIED
     ONLY IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE AND WITH THE INCLUSION
     OF THE ABOVE COPYRIGHT NOTICE.  THIS SOFTWARE OR ANY OTHER COPIES
     THEREOF MAY NOT BE PROVIDED OR OTHERWISE MADE AVAILABLE TO ANY OTHER
     PERSON.  NO TITLE TO AND OWNERSHIP OF THE SOFTWARE IS HEREBY TRANSFERRED.

     THE INFORMATION IN THIS SOFTWARE IS SUBJECT TO CHANGE WITHOUT NOTICE AND
     SHOULD NOT BE CONSTRUED AS A COMMITMENT BY HEWLETT-PACKARD COMPANY.

     HP ASSUMES NO RESPONSIBILITY FOR THE USE OR RELIABILITY OF ITS
     SOFTWARE ON EQUIPMENT THAT IS NOT SUPPLIED BY HP.

     NO RESPONSIBILITY IS ASSUMED FOR THE USE OR RELIABILITY OF SOFTWARE
     ON EQUIPMENT THAT IS NOT SUPPLIED BY HEWLETT-PACKARD COMPANY.

     SUPPORT FOR THIS SOFTWARE IS NOT COVERED UNDER ANY HP SOFTWARE
     PRODUCT SUPPORT CONTRACT, BUT MAY BE PROVIDED UNDER THE TERMS OF THE
     CONSULTING AGREEMENT UNDER WHICH THIS SOFTWARE WAS DEVELOPED.

#!/bin/ksh
function handle_signal
{
        print -n "pid $$ recieved $2 "
        if [[ $1 = 0 ]];then
            print but foreground command ended successfully
        else
                if [[ $1 = $3 ]];then
                    print and so did the last foreground command
                else
                    print -n "and the exit status of the last foreground "
                    print command was $1
                fi
        fi
        # Kill our process group and then ourselves with SIGTERM, giving a 
        # pid of 0 sends the signal to our process group. Killing the process
        # group should kill us as well, this assumes that SIGTERM is not 
        # handled by any process in the process group.
        #
        # This code could be replaced with an exit with an exit value that
        # would indicate what the problem was to the caller. That is replace 
        # these two lines with:
        #
        # exit $3
        #
        # or a specific exit code could be used.
        #
        kill -TERM 0
        kill -TERM $$
        }
OS=$(uname -a | awk '{print $1}')
if [[ "$OS" = "Linux" ]]; then
    offset=256
elif [[ ("$OS" = "HP-UX") || 
        ("$OS" = "SunOS") || 
        ("$OS" = "OSF1") ]]; then
   offset=128
fi
trap 'RC=$?; handle_signal $RC SIGINT $offset+2' INT
trap 'RC=$?; handle_signal $RC SIGQUIT $offset+3' QUIT
/bin/sleep 20
echo $?

DNS not working

Ping to an IP address works 

ping 74.125.136.103

but this doesn't 

ping www.google.com

Check resolv.conf 

cat /etc/resolv.conf
nameserver 95.130.132.17
nameserver 95.130.132.18

I had changed internet provider and forgot to update this. Just to set it to the router address and let that do the resolution 

nameserver 192.168.1.1

File descriptors

exec 3<> /tmp/foo  #open fd 3 for r/w
echo "test" >&3
exec 3>&- #close fd 3.

exec 3<> myfile.txt
while read line <&3
do {
  echo "$line"
  (( Lines++ ));                   #  Incremented values of this variable
                                   #+ accessible outside loop.
                                   #  No subshell, no problem.
}
done
exec 3>&-

echo "Number of lines read = $Lines"     # 8

Mmm. See our output and also tee it to a log file! 

#!/bin/bash

echo hello

if test -t 1; then
    # Stdout is a terminal.
    exec >log
else
    # Stdout is not a terminal.
    npipe=/tmp/$$.tmp
    trap "rm -f $npipe" EXIT
    mknod $npipe p
    tee <$npipe log &
    exec 1>&-
    exec 1>$npipe
fi

echo goodbye

Create new image with kvm

Ref: http://www.cyberciti.biz/faq/kvm-virtualization-in-redhat-centos-scientific-linux-6/
Build an empty space for a CentOS virtual machine 

qemu-img create -f qcow2 centos.img 12G

Tried creating image with 

sudo virt-install -n CentOS --description "Trying out CentOS" --ram=1024 --vcpus=1 --cpu host --hvm --cdrom /home/bey9at77/Downloads/c6-x86_64-20130910-1.qcow2 --graphics vnc --disk path=/var/lib/libvirt/images/centos.img,bus=virtio,size=10

gives error 

Starting install...
Allocating 'centos.img'                                                                                                                                                                                               |  10 GB     00:00     
ERROR    internal error Process exited while reading console log output: char device redirected to /dev/pts/1
qemu-kvm: -drive file=/home/bey9at77/Downloads/c6-x86_64-20130910-1.qcow2.bz2,if=none,media=cdrom,id=drive-ide0-1-0,readonly=on,format=raw: could not open disk image /home/bey9at77/Downloads/c6-x86_64-20130910-1.qcow2: Permission denied

rpm / yum commands

Install an RPM Package

rpm -ivh pidgin-2.7.9-5.el6.2.i686.rpm

-i - install
-v - verbose
-h - print progress hashes

Check dependencies of RPM Package before Installing

rpm -qpR BitTorrent-5.2.2-1-Python2.4.noarch.rpm

-q - query
-p - list package capabilities
-R - list dependent package capabilities

Install RPM Package with all dependencies

yum install BitTorrent-5.2.2-1-Python2.4.noarch.rpm

Install RPM Package with all dependencies (when RPM has been downloaded to local machine)

yum localinstall BitTorrent-5.2.2-1-Python2.4.noarch.rpm

Force Install a RPM Package without dependencies

Package will not work if dependencies are required 

rpm -ivh --nodeps BitTorrent-5.2.2-1-Python2.4.noarch.rpm

Check an Installed RPM Package

rpm -q BitTorrent

List all files of an installed RPM package

rpm -ql BitTorrent

List All Installed RPM Packages

rpm -qa

Query information about an installed RPM package

rpm -qi vsftpd

Query information about a not yet installed RPM package

rpm -qip sqlbuddy-1.3.3-1.noarch.rpm

(Forcibly) Remove an RPM Package

Use package name (as seen in -qi above), not full name 

rpm -ev (--nodeps) vsftpd

Query a file that was installed as part of an RPM Package (which package contained this file)

rpm -qf /usr/bin/htpasswd

Verify an RPM package

Compares information of installed files of the package against the rpm database 

rpm -Vp sqlbuddy-1.3.3-1.noarch.rpm

Rebuild corrupted RPM database

cd /var/lib
rm __db*
rpm --rebuilddb
rpmdb_verify Packages

Install rpmforge repository

  • Download rpmforge-release-0.5.3-1.el6.rf.x86_64.rpm
  • Import the key
sudo rpm --import http://apt.sw.be/RPM-GPG-KEY.dag.txt
  • Install the repository
sudo rpm -i rpmforge-release-0.5.3-1.el6.rf.x86_64.rpm
  • Check the installation
rpm -K rpmforge-release-0.5.3-1.el6.rf.x86_64.rpm
  • Test it
sudo yum install terminator

Install rpmfusion repository

su -c 'yum localinstall --nogpgcheck http://download1.rpmfusion.org/free/el/updates/6/i386/rpmfusion-free-release-6-1.noarch.rpm http://download1.rpmfusion.org/nonfree/el/updates/6/i386/rpmfusion-nonfree-release-6-1.noarch.rpm'

config file for yum

Checkout this file for global yum config 

/etc/sysconfig/yum-cron-background

Setup Oracle Enterprise Linux (RedHat) with yum server

You need to download the yum .repo file from the server, as per the steps below. After this, you need to enable a flag in the .repo file as per your operating system version. Having done these two steps, when you run yum install <pkgname> command on your linux box, the Oracle's yum server will be scanned, the dependent & the relevant rpm's will be download and installed for you.

cd /etc/yum.repos.d

To download files here 

wget http://public-yum.oracle.com/public-yum-el5.repo

A file named public-yum-el5.repo will be created in your directory
Edit this file and enter enabled=1 against the operating systems which is relevant to you 

vi public-yum-el5.repo

Next run the yum command 

yum install package-name

To change to static IP address (Raspberry Pi)

As root: 

cd /etc/networks
vi interfaces

replace the line “iface eth0 inet dhcp” with 

iface eth0 inet static
address 192.168.1.100
netmask 255.255.255.0
gateway 192.168.1.1

You should also take a look at the file /etc/resolv.conf and check it has a nameserver entry (probably pointing at your default gateway) or direct to your ISP name servers. 

nameserver 192.168.1.1

Troubleshoot wireless network problems

Short summary of all the things you need to do in just few lines 

root@kali:~# iw dev
root@kali:~# ip link set wlan0 up
root@kali:~# iw wlan0 scan
root@kali:~# wpa_passphrase blackMOREOps >> /etc/wpa_supplicant.conf
root@kali:~# wpa_supplicant -i wlan0 -c /etc/wpa_supplicant.conf
root@kali:~# iw wlan0 link
root@kali:~# dhclient wlan0
root@kali:~# ping 8.8.8.8
(Where wlan0 is wifi adapter and blackMOREOps is SSID)
(Add Routing manually)
root@kali:~# ip route add default via 10.0.0.138 dev wlan0

To change to static IP address (Redhat/CentOS)

As root: 

vi /etc/sysconfig/network-scripts/ifcfg-eth0
DEVICE=eth0
BOOTPROTO=STATIC
IPADDR=192.168.1.111
NETMASK=255.255.255.0
GATEWAY=192.168.1.1
ONBOOT=yes
DNS1=8.8.8.8
DNS2=8.8.4.4

Resrart the network interface 

/etc/init.d/network stop
/etc/init.d/network start
or
service network restart

Check name server entry in resolv.conf 

vi /etc/resolv.conf
nameserver 192.168.1.1

Enable processes / services to start at boot time

sudo chkconfig httpd on
sudo chkconfig mysqld on

Run a command on lots of servers in parallel

This is a damn fine AIX utility - part of the CSM Distributed Shell. 

dsh -a "ls -al /etc/apache2/*conf"

will list the Apache configuration file on all reachable servers (nodes)

Download a gzip file and pipe it into tar

cd ~ && wget -O - "https://www.dropbox.com/download?plat=lnx.x86_64" | tar xzf -

Check in a script to make sure it is run only by root

RUID=`/usr/bin/id|$AWK -F\( '{print $1}'|$AWK -F\= '{print $2}'`
if [ ${RUID} != "0" ];then
   $ECHO "This script must be executed as root"
   exit 1
fi

Set terminal to use Backspace key to erase previous character instead of Control-H

Been looking for this for a long time.
You can put: 

stty erase <CTRL-V><Backspace key>

in your .profile but this will be ruined if you do a copy/paste into another file.
I wanted a way of doing this without entering the control character in the .profile. Finally stumbled upon it. And it's so simple. Just escape the caret! 

stty erase \^?

Put this in the .profile. It's copy/pastable and it works!
If you want CTRL-H to be your erase character, just do this: 

stty erase \^H

Play with the terminal settings and reset them again either side of requesting a password

The -g option of stty gives a compact list of all the settings or the terminal and can be used as input to stty 

OLDCONFIG=`stty -g`          # save terminal configuration
stty -echo                   # turn character echoing off
echo "Enter password: \c"
read PASSWD                  # get the password
stty $OLDCONFIG              # restore terminal configuration

Reset terminal to "sane" characteristics

If you've done a cat of a binary file or something else weird and your terminal is left in a mess, the following key sequence should bring it back to normal 

<CTRL-J>stty sane<CTRL-J>

Install OpenOffice on RedHat Enterprise when yum install doesn't!

Download Package 

wget http://sourceforge.net/projects/openofficeorg.mirror/files/4.0.1/binaries/en-US/Apache_OpenOffice_4.0.1_Linux_x86-64_install-rpm_en-US.tar.gz/download -O Apache_OpenOffice_4.0.1_Linux_x86-64_install-rpm_en-US.tar.gz

Change to root 

sudo su -

Remove the old stuff 

yum remove openoffice* libreoffice*

Extract Package 

tar -xvf Apache_OpenOffice_4.0.1*
cd en-US

Install Package and exit root 

rpm -Uvh RPMS/*.rpm RPMS/desktop-integration/openoffice4.0-redhat-*.rpm
exit

Start it 

openoffice4

What does this do?

while IFS= read -r line; do
    echo "[$(date "+%F %T")] - $line"
done < <(iwevent)

or 

while IFS= read -r line; do
    printf "%s\n%s\n" "$line" "Yohooo! One more package."
done < <(tcpdump -i any -nS)

List of Special Characters and what they mean

From Bruce Barnett 

Character	Where	Meaning
<RETURN>	csh, sh	Execute command
#	csh, sh, ASCII files	Start a comment
<SPACE>	csh, sh	Argument separator
`	csh, sh	Command substitution
"	csh, sh	Weak Quotes
'	csh, sh	Strong Quotes
\	csh, sh	Single Character Quote
variable	sh, csh	Variable
variable	csh, sh	Same as variable
|	csh, sh	Pipe character
^	sh	Pipe Character
&	csh, sh	Run program in background
?	csh, sh	Match one character
*	csh, sh	Match any number of characters
;	csh, sh	Command separator
;;	sh	End of Case statement
~	csh	Home Directory
~user	csh	User's Home Directory
!	csh	History of Commands
-	Programs	Start of optional argument
$#	csh, sh	Number of arguments to script
$*	csh, sh	Arguments to script
$@	sh	Original arguments to script
$-	sh	Flags passed to shell
$?	sh	Status of previous command
$$	sh	Process identification number
$!	sh	PID of last background job
&&	sh	Short-circuit AND
||	sh	Short-circuit OR
.	csh, sh	Typ. filename extension
.	sh	Source a file and execute as command
:	sh	Nothing command
:	sh	Separates Values in environment variables
:	csh	Variable modifier
Character	Where	Meaning
[ ]	csh, sh	Match range of characters
[ ]	sh	Test
%job	csh	Identifies job Number
(cmd;cmd)	csh. sh	Runs cmd;cmd as a sub-shell
{ }	csh	In-line expansions
{cmd;cmd }	sh	Like (cmd;cmd ) without a subshell
>ofile	csh, sh	Standard output
>>ofile	csh, sh	Append to standard output
<ifile	csh, sh	Standard Input
<<word	csh, sh	Read until word, substitute variables
<<\word	csh, sh	Read until word, no substitution
<<-word	sh	Read until word, ignoring TABS
>>!file	csh	Append to file, ignore error if not there
>!file	csh	Output to new file, ignore error if not there
>&file	csh	Send standard & error output to file
<&digit	sh	Switch Standard Input to file
<&-	sh	Close Standard Input
>&digit	sh	Switch Standard Output to file
>&-	sh	Close Standard Output
digit1<&digit2	sh	Connect digit2 to digit1
digit<&-	sh	Close file digit
digit2>&digit1	sh	Connect digit2 to digit1
digit>&-	sh	Close file digit
Retrieved from "https://mailnest.com/dbawiki/index.php?title=Unix/Linux&oldid=3398"