6. Swap Management ( Real Time)

6. Swap management

Issue 1: the server went down or hangs.

A: go to console, check the error messages, and reboot

Issue 2: why the server slower down?

A:

1. # vmstat

It checks the page in & page out size.

Page in should be more size than page out.

If page out size is more, then it means page out didn’t release memory to swap.

Then

# ps –ef | grep 1255

If root

# kill -9 1255

# kill –HUP 1255 (HUP for refresh)

Note: In root mostly java or oracle process runs, so kill that.

Note : if other applications inform them

# top (or) # prstat –a

It shows max memory occupied application. Then submit/inform report to the desired applications. Mostly /oracle, /java applications

If it is ‘root’ issue then delete the old files & compress the unuse files

To display volume table of content (vtoc)

# prtvtoc /dev/rdsk/c0d0s2

To display only CPU utilization:

# prstat –s CPU –n 10

5. CPU Management (Real Time)

5. CPU management

Show process of CPU, memory, and swap utilization

#top

Or

#prstat -a

Note: only ‘root’ issues we take care & other applications issues we will transfer /inform to them only.

(# ftp <file name>: sending to oracle/other applications)

To display volume table of content (vtoc)

# prtvtoc /dev/rdsk/c0d0s2

To display only CPU utilization:

# prstat –s CPU –n 10

4. User Management ( Real Time)

4. User Management

# id      (find out the user id information)

# dispuid        

# dispgid

# logins

# cat /etc/default/fs

Local = ufs

#listusers         (show no.of users)

Note: for user, there should be ‘home directory’.

Issues in user management:

ü  User permission

ü  User home directory missing

ü  Change the ownership (chown)

ü  Assign the permission (chmod)

ü  User missing in shadow file (pwconv)

User environment variable:

1. local.cshrc
2. local.login
3. local.profile

All file exists in /etc/skel file

# logins –l raj

How to assign groupid to user?

# groupadd unix

#usermod –g unix raj  (-g : primary group name)

# logins –l raj

#id –a raj

# usermod –G dba,dwh raj

Note: if not specifying anything to create user, it means it stored in /home

To find defadduser:

# find / -name defadduser  -print

# login –xl pra

$ ls –ltr

raj

$exit

Changing the owner ship: before it was raj, now I am changing as root

#chown root:other /sai/raj/aa

# ls –l /sai/raj/aa

#chown root:dba /sai/raj/aa

# groups

# dispgid

# cat /etc/group

Sol:143:-----     (3^rd one,if secondary group (-G) only it shows)

‘#userdel’ is a command to del the user but it cant delete the user home directories

#userdel –r:     command delete the home directory

To change the user name;

#usermod –l  roja raj

To change the group name:

#groupmod –n java sol

# man groupadd

# groupadd  ---

Above two, it show the options

#lf –F

#useradd –D (if it fire then only defadduser updates)

#cat /usr/sadm/defadduser

Issue1:

User entry missing in shadow

A: sync from passwd to shadow (pwconv)

# pwconv

Note: we can not sync from shadow to password.

Issue 2:

Password break

A: go to filesafe mode

# cd /a

#vi etc/shadow

    root/x/                 (remove 2^nd part i.e  x)

in system1:

# dtrace           /it trace all the system calls

In system2:

# rm     // rm commands not working

Now in system 1 it shows error message, it trace why ‘rm’ not working

#passwd –n 2 –w 3 –x 10 raj

User id creation: LDAP – secure – advance of NIS+

Configuration: implementation team take care

Note :

1. Once ticket comes for production server till 5 days no need to do anything.
2. Application & DBA both should be shutdown for rebooting the server

Issue 3: home directory missing

# useradd –u 144 –g dba –d /sai/ram –m ram

# usermod –d /opt/ram –m ram           (b4 ram was in sai)

# cd /opt

# ls

ram (now in /opt from /sai, in /sai not there)

issue 4: chage the ownership

# ls –ld /opt/ram

I want to change the owner ship from ram to raj

#ls –ltr

Ram dba

I want raj to access ram

#chown raj:dba /opt/ram/aa

#ls –ld /opt/ram/aa

Issue 5: only one user not loging in

ü  Check the permissions

ü  Check out the password, user

ü  Check password status, grep | username

Issue 6: one user not accessing other file

ü  Check user permissions

ü  File level permissions

Note: its not day to day activity depends on requirement

Issue 7:not a single user login to server

Then it is a server problem

Issue 8: if shadow file missing

A: sync

#pwconv

Issue 9: if passwd/opasswd file missing

A: Reconstruct

3. SMF (Service Managment Facility) - Real Time

3. SMF (Service Management Facility)

SMF:

ü  It is a advanced feature in Solaris10

ü  Which will control services under SMF

ü  We can start and stop the services

-          All the services stored in ‘/etc/init.d’

-          We can’t  stop services dependency/ dependent

#svcs               (display online & offline and legacy)

#svcs –a           (display online, offline, and disable)

#svcs |more

How to find out dependency?

#svcs –d svc:/network/nfs/server:default

How to find out dependent?

#svcs –D svc:/network/nfs/server:default

Note: if any one of these services in disable then NFS will not get into online

Issue 1: service (NFS) is in disable, but it is not getting ‘online’ why?

A: check out the dependency & dependent and make it online. Then make the service online. If it is in maintenance mode means configuration problem.  Configuration problem means implementation team take care.

Issue 2: why the service group not running? What is the reason?

#svcs –xv svc:/network/rpc/keyserv: default

Same above reason find out dependency and dependent, if any in disable mode make it online.

Issue 3: If service group lost

# cd /etc/svc

Repository.db

Service ‘repository.db’ file store in /etc/svc

Restoring ‘repository.db’

# cd /lib/svc/bin

# ./restore_repository              (execute)

To see execute or text

# file restore_repository

# cd /var/svc/profile

All SMF scripts are store under /var/svc/profile

Service categories:

#cd manifest

#ls

1. Application
2. milestone –runlevel information
3. platform
4. system
5. device
6. network
7. site

2. Booting Procedure (Real Time)

2. Boot Procedure

1. Boot prom:

ü  Prom runs the POST

ü  Search for the bootable devices

ü  Read the boot block

ü  Load the boot block

2. Boot program:

ü  Load the ufs boot block

ü  Try to load the secondary boot block

ü  If boot block corrupted

 # cd /usr/platform/’uname –i ‘/lib/fs/ufs

3. Kernel :

ü  Read the etc/system file, where etc/sytem has total kernel information

ü  It initializes all the modules itself

4. Init :

ü  Read  the etc/init file

ü  Total run level information]

ü  Svcstatd deamons runs

Issue 1: If boot block corrupted what you do?

A: run the following command and restore boot

 # cd /usr/platform/’uname –i ‘/lib/fs/ufs

1. File System Management (Real Time)

1. File System Management

File System issues are not a regular task.

Basic commands:

# uname –a                              To display the OS name, version, kernel version

# isainfo –b                             Display the kernel bit

# format                                  To find the available disk

# newfs –N                             Display the super block & primary block

Last + found                           consist of inode, data block information

# ls –s *                                   List all the details of files & directories

                                                (for touch 0Mb, for cat 2 blocks)

Q: How to check which file occupying more memory in File system?

A:

# cd /sai                       /sai is the mount point

# du –sh * | sort –nr                // displays the files occupying more memory

Q: Can we increase the file system in Solaris?

A: Generally we cannot increase the File System, but we can increase the File system if continuous cylinders available.

# growfs –M /sai /dev/rdsk/c0d0s6

# df –h /sai

Q: Can we decrease the File System?

A: No way, not possible

Note: to over come this problem SVM (Solaris Volume Manager) use.

How many are using the File System?

# cd /sai

# fuser –u /sai                          // to display how many users using File System

To check the minfree:

#fstyp –v /dev/rdsk/c0d0s0 | grep –i minfree

We can write the script for repeated loop of status:

# while true

> do

> fstyp –v /dev/rdsk/c0d0s0 |grep –i  minfree

> done

To increase the minfree:

# tunefs –m 20 /dev/rdsk/c0d0s0

To decrease the minfree:

# tunefs –m 10 /dev/rdsk/c0d0s0

Issue 1:

If file system exceeds the threshold value then ticket will generate automatically.

A:To resolve this

Check out more size files in File System

Compress the files or delete the old files, if it is so important consult the clients then delete.

Copy /keep the data:

# cp –r /usr/* .

To see old files;

# ls –ltr

To gip:

# gzip aa

To unzip:

# gunzip aa.gz

^*** To find top 10 files occupied more size:

# find . –xdev –type f –ls | sort –nr +6|head -10

^*** To find top 20 files occupied more size:

# find . –xdev –type f –ls | sort –nr +6|head -20

To move the data:

# mv /etc/aa /opt               move /aa to /opt

Issue 2:

If inodes are corrupted (some times file system also will not work)

A: To resolve:

I will ask the backup team to take backup of file system, & restore it freshly.

Issue 3:

Suppose my File System is 2 GB, its usage is 1GB, but it is showing 100%. Actually it has to show 50%.

A:To resolve:

Simply I unmount & mount the filesystem

Or

If it is not release the space, I will reboot. Then it will show 50%.

Issue 4:

User can not accessing the file system

A: To resolve:

Check out the file permissions, and give the proper permissions to that file.

# mount: mount file system from mnttab

# mountall: mount all file system from vfstab