OS Level Checks

This is very important to understand Os level concepts while working on SAP systems .What does different
processes mean when system is up and what it signifies.It helps lot while you are doing troubleshooting .Along with
this administrator must know what are different profile files available and their starting sequence.Let me Introduce you
one by one about different files and processes available at OS level:

# Important work processes at OS level :

You can check all processes on OS level by following command :

>ps -ef|grep <SID>adm
1- DW: this shows all work processes including dispatcher for one particual instance .If you have defined
n(dialog+background+spool+update+enq) number of processes in instance profile ,You will see n+1 number of dw
processes at OS level .So now questions comes why n+1 dw processes will be available at OS level .Answer is one
process stands for dispatcher work process.

2- ICMAN* : ICM process starts with ICMAN

3- GWRD: gateway process starts with naming convention

4- MS: Message server starts with this naming convention

#Overview of SAP Profile Files :

1- Start Profile :

Processes to be started .

Definition :

When you start an SAP instance on a host, the start profile defines which SAP services are started (message server,
dialog, gateway or enqueue process. for example). The
startsap program is responsible for starting these service processes, and it uses a start profile to begin
the startup process.

The processes that can be started include:

Application server
Message server
SNA Gateway
System log send demon
System log receive demon
Apart from the general profile parameters, such as the name of the SAP System (SAPSYSTEMNAME ),
instance number (SAPSYSTEM ) and name of the SAP instance ( INSTANCE_NAME ), the only parameter
names that are permitted in a start profile are:

Execute_xx (xx = 00-99)

To start operating system commands, which prepare the SAP System start. For example, you can use
this parameter to start the SAP-related database or to set up links to executables on UNIX platforms.

Start_Program_xx (xx = 00-99)

: To start an SAP instance, for example, on an application server.

Stop_Program_xx (xx = 00-99)

 To start an operating system command or SAP program after the SAP instance was stopped. For
example, the halting or removal of shared memory areas that were used by the SAP System.

The number

xx defines the execution sequence. The programs specified in Execute_ parameters are executed
before the programs listed in the Start_Program parameters. After the SAP instance has been
stopped, the programs specified in the Stop_Program parameters are started. Here is an example of a
start profile used to start a message server, an application server and an SNA gateway
Above definition Source : http://help.sap.com

*** In start sequence of SAP ,Start profile is the first profile to be started

2- DEFAULT Profile :

All Global parameters that should apply to all instances always defined here .So if you want to set any value for all
instances then always set that parameter in DEFAULT.PFL .
*** In starting sequence of SAP it is always read at second position means just after start profile ***

3- Instance Profile :
Parameters what apply to one instance always defined here .Like number of work processes etc
*** Last profile to be read while starting SAP system ***

From above sequence of profile read we can conclude that you can indenpendtly start and stop your dialog
instances without affecting your other instances .

# In work directory of SAP system we see following files:

1- stderr1....m : standard error file of program SAPSTART

2-sapstart<m>.trc : Trace files of program SAPSTART
3- sapstart.log : Startup log of SAPSTART
4-Dev_ms: trace file os message server
5-dev_W0..Wn: Trace files of work processes
6-Dev_rd: trace file os gateway work process
7:dev_icm : trace file of ICM
Unix Commands for SAP BASIS CONSULTANT
To handle day to day Basis activities ,Basis consultants need to know operating system commands effectively
.Handling windows operating system is very easy but when you are working on UNIX .You must very well aware
about all commands and their way of utilisation .

Manual command.

man man This is help command, and will explains you about online manual pages you can also use man in
conjunction with any command to learn more about that command for example.

man ls will explain about the ls command and how you can use it.

man -k pattern command will search for the pattern in given command.

Banner command.

banner prints characters in a sort of ascii art poster, for example to print wait in big letters. I will type

banner wait at unix command line or in my script. This is how it will look.

# # ## # ##### # # # # # ## # # # # # # ## # ###### # #
## ## # # # ## # # # # #

Cal command

cal command will print the calender on current month by default. If you want to print calander of august of 1965.
That's eightht month of 1965.

cal 8 1965 will print following results.

August 1965 S M Tu W Th F S 1 2 3 4 5 6 7 8 9 10 11 12 13 1415 16 17 18 19 20 2122 23 24 25 26 27

2829 30 31

Clear command

clear command clears the screen and puts cursor at beginning of first line.

Calendar command
calendar command reads your calendar file and displays only lines with current day.

For example in your calendar file if you have this

12/20 Test new software.1/15 Test newly developed 3270 product.1/20 Install memory on HP 9000 machine.

On dec 20th the first line will be displayed. you can use this command with your crontab file or in your login files.

Nohup command.

nohup command if added in front of any command will continue running the command or process even if you shut
down your terminal or close your session to machine. For exmaple, if I want to run a job that takes lot of time and
must be run from terminal and is called update_entries_tonight .

nohup update_entries_tonight will run the job even if terminal is shut down in middle of this job.

Tty command

Tty command will display your terminal. Syntax is

tty options

Options

-l will print the synchronous line number.

-s will return only the codes: 0 (a terminal), 1 (not a terminal), 2 (invalid options) (good for scripts)

File Management commands.

cat,cd, cp, file,head,tail, ln,ls,mkdir ,more,mv, pwd, rcp,rm, rmdir, wc.

Pwd

pwd command will print your home directory on screen, pwd means print working directory.

/usr/sap/ABC
is output for the command when I use pwd in /usr/sap/ABC directory.

Ls :

ls command is most widely used command and it displays the contents of directory.

options

ls will list all the files in your home directory, this command has many options.

ls -l will list all the file names, permissions, group, etc in long format.

ls -a will list all the files including hidden files that start with . .

ls -lt will list all files names based on the time of creation, newer files bring first.

ls Fx will list files and directory names will be followed by slash.

ls R will lists all the files and files in the all the directories, recursively.

ls -R | more will list all the files and files in all the directories, one page at a time.

Mkdir.

mkdir ABC will create new directory, i.e. here ABC directory is created.

cd:

cd ABC will change directory from current directory to ABC directory.

Use pwd to check your current directory and ls to see if ABC directory is there or not.

You can then use cd ABC to change the directory to this new directory.

Cat :

cat cal.txt cat command displays the contents of a file here cal.txt on screen (or standard out).
Head :

head filename by default will display the first 10 lines of a file.

If you want first 50 lines you can use head -50 filename or for 37 lines head -37 filename and so forth.

Tail :

tail filename by default will display the last 10 lines of a file.

If you want last 50 lines then you can use tail -50 filename.

More:

more command will display a page at a time and then wait for input which is spacebar. For example if you have a file
which is 500 lines and you want to read it all. So you can use

more filename

Wc :

wc command counts the characters, words or lines in a file depending upon the option.

Options

wc -l filename will print total number of lines in a file.

wc -w filename will print total number of words in a file.

wc -c filename will print total number of characters in a file.

File :

File command displays about the contents of a given file, whether it is a text (Ascii) or binary file. To use it type

file filename. For example I have cal.txt which has ascii characters about calander of current month and I have
resume1.doc file which is a binariy file in microsoft word. I will get

file resume.doc
resume1.doc: data

file cal.txt

cal.txt: ascii text

Cp :

cp command copies a file. If I want to copy a file named oldfile in a current directory to a file named newfile in a
current directory.

cp oldfile newfile

If I want to copy oldfile to other directory for example /tmp then

cp oldfile /tmp/newfile. Useful options available with cp are -p and -r . -p options preserves the modification time and
permissions, -r recursively copy a directory and its files, duplicating the tree structure.

Rcp:

rcp command will copy files between two unix systems and works just like cp command (-p and -i options too).

For example you are on a unix system that is called XYZ and want to copy a file which is in current directory to a
system that is called RAM in /usr/sap/ directory then you can use rcp command

rcp filename RAM:/usr/sap

You will also need permissions between the two machines. For more infor type man rcp at command line.

Mv:

mv command is used to move a file from one directory to another directory or to rename a file.

Some examples:

mv oldfile newfile will rename oldfile to newfile.

mv -i oldfile newfile for confirmation prompt.

mv -f oldfile newfile will force the rename even if target file exists.

mv * /usr/sap/ will move all the files in current directory to /usr/sap directory.

Ln :

Instead of copying you can also make links to existing files using ln command.

If you want to create a link to a file called abcfile in /usr/local/bin directory then you can enter this command.
ln myfile /usr/local/bin/abcfile

Some examples:

ln -s fileone filetwo will create a symbolic link and can exist across machines.

ln -n option will not overwrite existing files.

ln -f will force the link to occur.

Rm :.

To delete files use rm command.

Options:

rm oldfile will delete file named oldfile.

rm -f option will remove write-protected files without prompting.

rm -r option will delete the entire directory as well as all the subdirectories, very dangerous command.

Rmdir:

rmdir command will remove directory or directories if a directory is empty.

Options:

rm -r directory_name will remove all files even if directory is not empty.

rmdir ABC is how you use it to remove ABC directory.

rmdir -p will remove directories and any parent directories that are empty.

rmdir -s will suppress standard error messages caused by -p.

Comparison and Searching commands :

diff,dircmp, cmp, grep, find.

Diff :

diff command will compare the two files and print out the differences between.

Diff ABC ABC1

Above command will show all difference of one and two file

Cmp:

cmp command compares the two files. For exmaple I have two different files fileone and filetwo.

cmp fileone filetwo will give me

fileone filetwo differ: char 80, line 4

if I run cmp command on similar files nothing is returned.

-s command can be used to return exit codes. i.e. return 0 if files are identical, 1 if files are different, 2 if files are
inaccessible.

This following command prints a message 'no changes' if files are same

cmp -s fileone file1 && echo 'no changes' .

no changes

Dircmp:

dircmp command compares two directories. If i have two directories in my home directory named

dirone and dirtwo and each has 5-10 files in it. Then

dircmp dirone dirtwo will return this

Grep:

grep command is the most useful search command. You can use it to find processes running on system, to find a
pattern in a file, etc. It can be used to search one or more files to match an expression.

It can also be used in conjunction with other commands as in this following example, output of ps command is passed
to grep command, here it means search all processes in system and find the pattern sleep.

ps -ef | grep sleep will display all the sleep processes running in the system as follows.

Options:

-b option will precede each line with its block number.

-c option will only print the count of matched lines.
-i ignores uppercase and lowercase distinctions.
-l lists filenames but not matched lines.
other associated commands with grep are egrep and fgrep. egrep typically runs faster. for more information type man
egrep or man fgrep in your system.

Find :
Find command is a extremely useful command. you can search for any file anywhere using this command provided
that file and directory you are searching has read write attributes set to you ,your, group or all. Find descends
directory tree beginning at each pathname and finds the files that meet the specified conditions. Here are some
examples.

Some Examples:
find $HOME -print will lists all files in your home directory.

find /work -name chapter1 -print will list all files named chapter1 in /work directory.

find / -type d -name 'man*' -print will list all manpage directories.

find / -size 0 -ok rm {} \; will remove all empty files on system.

conditions of find

-atime +n |-n| n will find files that were last accessed more than n or less than -n days or n days.

-ctime +n or -n will find that were changed +n -n or n days ago.

-depth descend the directory structure, working on actual files first and then directories. You can use it with cpio
command.

-exec commad {} \; run the Unix command on each file matched by find. Very useful condition.

-print print or list to standard output (screen).

-name pattern find the pattern.

-perm nnnfind files whole permission flags match octal number nnn.

-size n find files that contain n blocks.

-type c Find file whole type is c. C could be b or block, c Character special file, d directory, p fifo or named pipe, l
symbolic link, or f plain file.

Text processing commands:

cut,paste, sort, uniq,awk,sed,vi.

Cut .
cut command selects a list of columns or fields from one or more files.
Option -c is for columns and -f for fields. It is entered as
cut options [files]
for example if a file named testfile contains
this is firstlinethis is secondlinethis is thirdline
Examples:
cut -c1,4 testfile will print this to standard output (screen)
tststs
It is printing columns 1 and 4 of this file which contains t and s (part of this).
Options:
-c list cut the column positions identified in list.
-f list will cut the fields identified in list.
-s could be used with -f to suppress lines without delimiters.

Paste :

paste command merge the lines of one or more files into vertical columns separated by a tab.

for example if a file named testfile contains

this is firstline
and a file named testfile2 contains
this is testfile2
then running this command
paste testfile testfile2 > outputfile
will put this into outputfile
this is firstline this is testfile2
it contains contents of both files in columns.
who | paste - - will list users in two columns.

Options:

-d'char' separate columns with char instead of a tab.

-s merge subsequent lines from one file.

Sort:

sort command sort the lines of a file or files, in alphabetical order. for example if you have a file named testfile with
these contents
zzzaaa1234yuerwerqwwwwe
Then running

sort testfile

will give us output of

1234aaaqwwwerwweyuerzzz
Options:
-b ignores leading spaces and tabs.
-c checks whether files are already sorted.
-d ignores punctuation.
-i ignores non-printing characters.
-n sorts in arithmetic order.
-ofile put output in a file.
+m[-m] skips n fields before sorting, and sort upto field position m.
-r reverse the order of sort.
-u identical lines in input file apear only one time in output.

Uniq:

uniq command removes duplicate adjacent lines from sorted file while sending one copy of each second file.
Examples

sort names | uniq -d will show which lines appear more than once in names file.

Options:
-c print each line once, counting instances of each.
-d print duplicate lines once, but no unique lines.
-u print only unique lines.

Awk and Nawk:

awk is more like a scripting language built in on all Unix systems. Although mostly used for text processing, etc.
Here are some examples which are connected with other commands.

Examples:
df -t | awk 'BEGIN {tot=0} $2 == "total" {tot=tot+$1} END {print (tot*512)/1000000}' Will give total space in your system
in megabytes.
Here the output of command df -t is being passed into awk which is counting the field 1 after pattern "total" appears.
Same way if you change $1 to $4 it will accumulate and display the addition of field 4

which is used space.

for more information about awk and nawk command in your system enter man awk or man nawk.

Sed :

sed command launches a stream line editor which you can use at command line.
you can enter your sed commands in a file and then using -f option edit your text file. It works as

sed [options] files

options:
-e 'instruction' Apply the editing instruction to the files.
-f script Apply the set of instructions from the editing script.
-n suppress default output.
for more information about sed, enter man sed at command line in your system.

Vi editor.

vi command launches a visual editor. To edit a file type

vi filename

vi editor is a default editor of all Unix systems. It has several modes. In order to write characters you will need to hit i
to be in insert mode and then start typing. Make sure that your terminal has correct settings, vt100 emulation works
good if you are logged in using pc.

Once you are done typing then to be in command mode where you can write/search/ you need to hit :w filename to
write

and in case you are done writing and want to exit

:w! will write and exit.
options:

i for insert mode.

I inserts text at the curson

A appends text at the end of the line.
a appends text after cursor.
O open a new line of text above the curson.
o open a new line of text below the curson.
: for command mode.
<escape> to invoke command mode from insert mode.
:!sh to run unix commands.
x to delete a single character.
dd to delete an entire line
ndd to delete n number of lines.
d$ to delete from cursor to end of line.
yy to copy a line to buffer.
P to paste text from buffer.
nyy copy n number of lines to buffer.
:%s/stringA/stringb /g to replace stringA with stringB in whole file.
G to go to last line in file.
1G to go to the first line in file.
w to move forward to next word.
b to move backwards to next word.
$ to move to the end of line.
J join a line with the one below it.
/string to search string in file.
n to search for next occurence of string.