Thursday, 10 May 2012

UNIX - LINUX Interview Questions and Answers :

UNIX - LINUX Interview Questions and Answers :

1. How are devices represented in UNIX?

All devices are represented by files called special files that are located in/dev directory. Thus, device files and other files are named and accessed in the same way. A 'regular file' is just an ordinary data file in the disk. A 'block special file' represents a device with characteristics similar to a disk (data transfer in terms of blocks). A 'character special file' represents a device with characteristics similar to a keyboard (data transfer is by stream of bits in sequential order).

2. What is 'inode'?

All UNIX files have its description stored in a structure called 'inode'. The inode contains info about the file-size, its location, time of last access, time of last modification, permission and so on. Directories are also represented as files and have an associated inode. In addition to descriptions about the file, the inode contains pointers to the data blocks of the file. If the file is large, inode has indirect pointer to a block of pointers to additional data blocks (this further aggregates for larger files). A block is typically 8k.

Inode consists of the following fields:
File owner identifier
File type
File access permissions
File access times
Number of links
File size
Location of the file data

3. Brief about the directory representation in UNIX

A Unix directory is a file containing a correspondence between filenames and inodes. A directory is a special file that the kernel maintains. Only kernel modifies directories, but processes can read directories. The contents of a directory are a list of filename and inode number pairs. When new directories are created, kernel makes two entries named '.' (refers to the directory itself) and '..' (refers to parent directory).

System call for creating directory is mkdir (pathname, mode).

4. What are the Unix system calls for I/O?

open(pathname,flag,mode) - open file
creat(pathname,mode) - create file
close(filedes) - close an open file
read(filedes,buffer,bytes) - read data from an open file
write(filedes,buffer,bytes) - write data to an open file
lseek(filedes,offset,from) - position an open file
dup(filedes) - duplicate an existing file descriptor
dup2(oldfd,newfd) - duplicate to a desired file descriptor
fcntl(filedes,cmd,arg) - change properties of an open file
ioctl(filedes,request,arg) - change the behaviour of an open file

The difference between fcntl anf ioctl is that the former is intended for any open file, while the latter is for device-specific operations.

5. How do you change File Access Permissions?

Every file has following attributes:

owner's user ID ( 16 bit integer )

owner's group ID ( 16 bit integer )

File access mode word

'r w x -r w x- r w x'

(user permission-group permission-others permission)

r-read, w-write, x-execute

To change the access mode, we use chmod(filename,mode).

Example 1:

To change mode of myfile to 'rw-rw-r–' (ie. read, write permission for user - read,write permission for group - only read permission for others) we give the args as:

chmod(myfile,0664) .

Each operation is represented by discrete values

'r' is 4

'w' is 2

'x' is 1

Therefore, for 'rw' the value is 6(4+2).

Example 2:

To change mode of myfile to 'rwxr–r–' we give the args as:


6. What are links and symbolic links in UNIX file system?

A link is a second name (not a file) for a file. Links can be used to assign more than one name to a file, but cannot be used to assign a directory more than one name or link filenames on different computers.

Symbolic link 'is' a file that only contains the name of another file.Operation on the symbolic link is directed to the file pointed by the it.Both the limitations of links are eliminated in symbolic links.

Commands for linking files are:

Link ln filename1 filename2

Symbolic link ln -s filename1 filename2

7. What is a FIFO?

FIFO are otherwise called as 'named pipes'. FIFO (first-in-first-out) is a special file which is said to be data transient. Once data is read from named pipe, it cannot be read again. Also, data can be read only in the order written. It is used in interprocess communication where a process writes to one end of the pipe (producer) and the other reads from the other end (consumer).

8. How do you create special files like named pipes and device files?

The system call mknod creates special files in the following sequence.

1. kernel assigns new inode,

2. sets the file type to indicate that the file is a pipe, directory or special file,

3. If it is a device file, it makes the other entries like major, minor device numbers.

For example:

If the device is a disk, major device number refers to the disk controller and minor device number is the disk.

9. Discuss the mount and unmount system calls

The privileged mount system call is used to attach a file system to a directory of another file system; the unmount system call detaches a file system. When you mount another file system on to your directory, you are essentially splicing one directory tree onto a branch in another directory tree. The first argument to mount call is the mount point, that is , a directory in the current file naming system. The second argument is the file system to mount to that point. When you insert a cdrom to your unix system's drive, the file system in the cdrom automatically mounts to /dev/cdrom in your system.

10. How does the inode map to data block of a file?

Inode has 13 block addresses. The first 10 are direct block addresses of the first 10 data blocks in the file. The 11th address points to a one-level index block. The 12th address points to a two-level (double in-direction) index block. The 13th address points to a three-level(triple in-direction)index block. This provides a very large maximum file size with efficient access to large files, but also small files are accessed directly in one disk read.

11. What is a shell?

A shell is an interactive user interface to an operating system services that allows an user to enter commands as character strings or through a graphical user interface. The shell converts them to system calls to the OS or forks off a process to execute the command. System call results and other information from the OS are presented to the user through an interactive interface. Commonly used shells are sh,csh,ks etc.

12. Brief about the initial process sequence while the system boots up.

While booting, special process called the 'swapper' or 'scheduler' is created with Process-ID 0. The swapper manages memory allocation for processes and influences CPU allocation. The swapper inturn creates 3 children:
the process dispatcher,
vhand and

with IDs 1,2 and 3 respectively.

This is done by executing the file /etc/init. Process dispatcher gives birth to the shell. Unix keeps track of all the processes in an internal data structure called the Process Table (listing command is ps -el).

13. What are various IDs associated with a process?

Unix identifies each process with a unique integer called ProcessID. The process that executes the request for creation of a process is called the 'parent process' whose PID is 'Parent Process ID'. Every process is associated with a particular user called the 'owner' who has privileges over the process. The identification for the user is 'UserID'. Owner is the user who executes the process. Process also has 'Effective User ID' which determines the access privileges for accessing resources like files.
getpid() -process id
getppid() -parent process id
getuid() -user id
geteuid() -effective user id

14. Explain fork() system call.

The `fork()' used to create a new process from an existing process. The new process is called the child process, and the existing process is called the parent. We can tell which is which by checking the return value from `fork()'. The parent gets the child's pid returned to him, but the child gets 0 returned to him.

15. Predict the output of the following program code




printf("Hello World!");



Hello World!Hello World!


The fork creates a child that is a duplicate of the parent process. The child begins from the fork().All the statements after the call to fork() will be executed twice.(once by the parent process and other by child). The statement before fork() is executed only by the parent process.

16. Predict the output of the following program code



fork(); fork(); fork();

printf("Hello World!");



"Hello World" will be printed 8 times.


2^n times where n is the number of calls to fork()

17. List the system calls used for process management:

System calls Description
fork() To create a new process
exec() To execute a new program in a process
wait() To wait until a created process completes its execution
exit() To exit from a process execution
getpid() To get a process identifier of the current process
getppid() To get parent process identifier
nice() To bias the existing priority of a process
brk() To increase/decrease the data segment size of a process.

18. How can you get/set an environment variable from a program?

Getting the value of an environment variable is done by using `getenv()'. Setting the value of an environment variable is done by using `putenv()'.

19. How can a parent and child process communicate?

A parent and child can communicate through any of the normal inter-process communication schemes (pipes, sockets, message queues, shared memory), but also have some special ways to communicate that take advantage of their relationship as a parent and child. One of the most obvious is that the parent can get the exit status of the child.

20. What is a zombie?

When a program forks and the child finishes before the parent, the kernel still keeps some of its information about the child in case the parent might need it - for example, the parent may need to check the child's exit status. To be able to get this information, the parent calls `wait()'; In the interval between the child terminating and the parent calling `wait()', the child is said to be a `zombie' (If you do `ps', the child will have a `Z' in its status field to indicate this.)

21. What are the process states in Unix?

As a process executes it changes state according to its circumstances. Unix processes have the following states:

Running : The process is either running or it is ready to run .

Waiting : The process is waiting for an event or for a resource.

Stopped : The process has been stopped, usually by receiving a signal.

Zombie : The process is dead but have not been removed from the process table.

22.What is LILO?

LILO stands for Linux boot loader. It will load the MBR, master boot record, into the memory, and tell the system which partition and hard drive to boot from.

23.What is the main advantage of creating links to a file instead of copies of the file?

A: The main advantage is not really that it saves disk space (though it does that too) but, rather, that a change of permissions on the file is applied to all the link access points. The link will show permissions of lrwxrwxrwx but that is for the link itself and not the access to the file to which the link points. Thus if you want to change the permissions for a command, such as su, you only have to do it on the original. With copies you have to find all of the copies and change permission on each of the copies.

24. Write a command to find all of the files which have been accessed within the last 30 days.

find / -type f -atime -30 > December.files

This command will find all the files under root, which is ‘/’, with file type is file. ‘-atime -30′ will give all the files accessed less than 30 days ago. And the output will put into a file call December.files.

25.What is the most graceful way to get to run level single user mode?

A: The most graceful way is to use the command init s.
If you want to shut everything down before going to single user mode then do init 0 first and from the ok prompt do a boot -s.

26. What does the following command line produce? Explain each aspect of this line.

$ (date ; ps -ef | awk ‘{print $1}’ | sort | uniq | wc -l ) >> Activity.log

A: First let’s dissect the line: The date gives the date and time as the first command of the line, this is followed by the a list of all running processes in long form with UIDs listed first, this is the ps -ef. These are fed into the awk which filters out all but the UIDs; these UIDs are piped into sort for no discernible reason and then onto uniq (now we see the reason for the sort - uniq only works on sorted data - if the list is A, B, A, then A, B, A will be the output of uniq, but if it’s A, A, B then A, B is the output) which produces only one copy of each UID.

These UIDs are fed into wc -l which counts the lines - in this case the number of distinct UIDs running processes on the system. Finally the results of these two commands, the date and the wc -l, are appended to the file "Activity.log". Now to answer the question as to what this command line produces. This writes the date and time into the file Activity.log together with the number of distinct users who have processes running on the system at that time. If the file already exists, then these items are appended to the file, otherwise the file is created.

Tuesday, 8 May 2012

Upgrade from PowerPath 5.1 to PowerPath 5.5 on AIX 5.3 TL12 SP1

I am going to talk about EMC PowerPath upgrade on AIX.

I have performed this upgrade from PowerPath 5.1 to PowerPath 5.5 on AIX 5.3 TL12 SP1.

1.    Disable PowerPath boot and auto varyon option for the datavg’s

pprootdev off
chvg –a n vgXX

2.    Select blank disk, and copy

alt_disk_copy -d {hdiskpowerXX}

3.    Identify the hdiskX to set the boot list manually. Choose the hdiskX which is available on dev # 0 by running powermt display dev=hdiskpowerXX

powermt display dev=hdiskpower31

Pseudo name=hdiskpower31
Invista ID=FNM00112400413
Logical device ID=6000144000000010E0595664D94DA740
state=alive; policy=ADaptive; priority=0; queued-IOs=0; ========================================================================
--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---
###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors
   1 fscsi1                   hdisk130  0E        active  alive       0      0
   2 fscsi2                   hdisk163  0B        active  alive       0      0
   2 fscsi2                   hdisk196  07        active  alive       0      0
   3 fscsi3                   hdisk229  06        active  alive       0      0
   3 fscsi3                   hdisk262  0A        active  alive       0      0
   0 fscsi0                   hdisk31   03        active  alive       0      0
   0 fscsi0                   hdisk64   0F        active  alive       0      0
   1 fscsi1                   hdisk97   02        active  alive       0      0

bootlist -m normal hdisk31
bootlist -m normal -o

4.    Reboot the server

shutdown -Fr

5.    After reboot, Verify data VGs status, make note of prior root disk, and cleanup

lsvg –o [ Verify the datavgs should not be active ] alt_rootvg_op -X old_rootvg

6.    Remove the devices the PowerPath

powermt remove dev=all

7.    Remove the datavg disks.

lspv | grep -wv rootvg | awk '{print $1'} | xargs -l rmdev -dl
lsdev –Cc disk [ Verify datavg disks are removed ] 

8.    Remove old EMC PowerPath for filesets.

installp -u "EMCpower*"

9.    Install new driver

mount nfs_server:/mount-point/5.5.3 /mnt

installp -acgXYd /mnt EMCpower.base

EMCpower.encryption \

lslpp –l “EMCpower*”

powermt check_registration [ verify the output ]

Product: PowerPath
Capabilities: All

powermt version [ verify the output ]

EMC powermt for PowerPath (c) Version 5.5 P 03 (build 1)\

10.   Run cfgmgr and check powermt output, then save the configuration.

cfgmgr -v
powermt config
powermt display dev=all
powermt save

11.   Enable auto varyon option for the datavgs and PowerPath boot

varyonvg vgXX
chvg –a y vgXX
pprootdev on

12.   Reboot the server and check the server comes up without any issues.

shutdown –Fr

13.   Copy current root disk back to original

pprootdev fix
alt_disk_copy –d hdiskpowerXX

14.   Set the boot list as per the process mentioned on step 3 and reboot the server with the original disk

bootlist –m normal hdiskX
shutdown –Fr

15.   Once the server is up with the original disk, Remove the alternate copy (new disk)

alt_rootvg_op –X old_rootvg

Twitter Bird Gadget