Linux Operation Essentials
March 27, 2017 · View on GitHub
Prev: Linux Essentials
Reading Operating System Data
cat /etc/system-release # get version of OS
lsb_release -d # another way to get version of OS
uname -r # get kernel version
cat /proc/version # another way to get kernel version
cat /proc/cmdline # arguments passed to kernal at boot
lsblk # list block devices
Starting and Stopping CentOS 7
Message user
# to send message to individual message
write tux
# or
cat > message <<END
The server is coming down at 5pm
Ensure you save and logoff
END
cat message
# to broadcast message to all users
wall < message
# To turn messages on/off
mesg y
mesg n
Shutdown commands
halt
poweroff
reboot
# legacy commands
init --help
telinit --help
#
shutdown -h 10 "the system is going down in 10min"
# to cancel:
ctrl-c
shutdown -c
# nologin file is create when time to shutdown is less than 5 minutes
cat /run/nologin
Changing Runlevels
who -r # show me my current run level - 5 being the graphical target
runlevel # ditto
systemctl get-default
systemctl set-default multi-user.target
# to change the run level while system is running
systemctl isolate multi-user.target # run level 3 - multi-user, same a graphical target level but without the graphics
systemctl isolate rescue.target # single user level 1, will lose network connections
Selecting Runlevels at Boot
- Start up virtual instance (must be powered off first)
- click e to edit the default kernel (default entry)
- scroll down to the line starting "linux16"
- navigate to end of line using ctrl-e and append:
- "1" or preferrably,
- "systemd.unit=rescue.target"
- ctrl-x (or F10) to exit and continue booting
The Boot Process
Managing GRUB Recovery
vi /etc/default/grub
# set GRUB_DISABLE_RECOVERY="false" to enable recovery mode
grub2-mkconfig -o /boot/grub2/grub.cfg
reboot
Recover Lost Root Passwords
- Reboot machine
- At boot option inout e at kernel entry to edit
- scroll down to the line starting "linux16"
- navigate to end of line using ctrl-e and:
- remove: rhgb quiet # so we can see the boot process
- append: rd.break enforcing=0
- ctrl-x
switch_root:/# mount -o remount,rw /sysroot # to remount root filesystem - currently set as readonly - need it read-right so we can set the password
chroot /sysroot # to set a false root which points through to the sysroot directory
# we are no working with real filesystem, which was previously mounted into sysroot
passwd # to reset password
# assuming no error...
exit # get out of chroot environment
# up arrow to retrieve mount command and edit
mount -o remount,ro /sysroot
exit # to exit and continue with boot process
# logon using the new password
restorecon /etc/shadow # since password was set outside of normal security context
setenforce 1 # selinux
Managing GRUB2
GRUB - GRand Unified Bootloader
Re-installing GRUB
# as root
grub2-install /dev/sda # install into the master boot record - /dev/sda on Centos 7
Manage GRUB2 Defaults
vi /etc/default/grub
# make changes and exit - then run
grub2-mkconfig -o /boot/grub2/grub.cfg
Manage GRUB2 with grubby
grubby --default-kernel
grubby --set-default /boot/vmlinuz-3.10.0-514.el7.x86_64
grubby --default-kernel
grubby --info=ALL
grubby --info /boot/vmlinuz-3.10.0-514.el7.x86_64
grubby --remove-args="rhgb quiet" --update-kernel !$
reboot
Password Protect GRUB2
Cleartext password
cp /etc/grub.d/01_users .
cd /etc/grub.d
vi 01_users # replace contents with the following (minus hash symbols)
# #!/bin/sh -e
# cat << EOF
# set superusers="andrew"
# password andrew L1nux
# EOF
grub2-mkconfig -o /boot/grub2/grub.cfg
reboot
# when we click _e_ to edit boot option this will now ask for a user and password
Encrypted password
grub2-mkpasswd-pbkdf2 # this will output a password hash - copy this to the clipboard
vi /etc/grub.d/01_users
# #!/bin/sh -e
# cat << EOF
# set superusers="andrew"
# password_pbkdf2 andrew grub.pbkdf2.sha512.10000.DDF4EAC30F5A9BB6809BF9E02C384E1B335CA469584CE6855069CE8FF991CDE85E114A18AA27D1EE9ED269F5B8C357A9CDFD886B03F4D80FB69565EBEE1F0A53.F7E4269A1746100DD4254B73AC9872A17EA0D1DAF1CFD282D9F99A99B13DC88C1B5579653831BA54894DCD3794C96405EFD646E52FF12D9A7A043C364F394B6C
# EOF
grub2-mkconfig -o /boot/grub2/grub.cfg
reboot
Custom GRUB2 Entries
vi /etc/grub.d/40_custom
# add details as required
grub2-mkconfig -o /boot/grub2/grub.cfg
reboot
Managing Linux Processes
Listing Processes with ps
ps
# BSD format
ps aux
# UNIX format
ps -f # full listing
ps -F # extra full listing
ps -l # long listing
ps -e # list everything
pstree
The /proc directory and the $$ variable
# $$ variable show the process id of the current process (the bash shell)
ps -p $$ -f
# double check
ps
echo $$
# we can see this also in the /proc directory
cd /proc
cd $$
Send Signals with kill
stty -a # to view keyboard shortcuts for sending signals
kill -l # to see the signals that can be sent to a process
kill -15 PID # terminate
kill -9 PID # signal cannot be caught so will always execute
Shortcuts with pgrep, pkill, and top
pgrep sshd # list process id of sshd processes
ps -F -p $(pgrep sshd)
# above is actually better than "ps -eF | grep sshd" as it includes headers and doesn't incude itself
# create some background sleep processes
sleep 100&
sleep 100&
sleep 100&
pgrep sleep # to list these
pkill sleep # to kill these
top
# "f" to show menu of what to sort by
# arrow up or down then "s" and escape to select
# "q" to quit
Process Priority
Backgrounding Tasks
sleep 1000&
jobs
sleep 1000 # without the ampersand
(ctrl-z) # to move task to background, suspended
jobs # this will show both our sleep jobs - the most recently one with a "+" symbol next to it - this means it has focus
bg # to resume the job in the background - the job that has focus
jobs
fg # to bring a currently in focus background job into the foreground
fg <nr> to bring job <nr> to the foreground
Configuring Process Priority using 'nice'
ps -l
# Columns PRI (priority) and NI (nice) should show 80 and 0 respectively for our sleep jobs
# Process Priority can be between 60 (high) and 99 (low) - default 80
# Nice can be between -20 and +19 - default 0
# manually set the nice value
nice -n 1 sleep 1000 &
renice -n 10 -p 2930 # where -p is PID
# only root can set priority lower than current value or nice value less than zero
# edit default values in the following file
vi /etc/security/limits.conf
# and add the following line
# tux - priority 10 # for a group add @ at start, or use * for all users
su - tux # to start a new shell
sleep 250&
ps -l # see new processes showing default priority of 10
Monitor Linux Performance
rpm -ql procps-ng # get listing of items included in procps-ng (next-generation package for monitoring processes)
rpm -ql procps-ng | grep '^/usr/bin/' # to see all 'programs' installed as part of this package
rpm -gf /usr/bin/ps # (-qf: query file) see that ps command is actually part of procps-ng package
pwdx and pmap commands
free -h # free/used memory
# process memory map
pmap $$
# report current working directory of a process
pwdx $$
pwdx $(pgrep sshd)
uptime and tload
uptime
lscpu # list cpu's
cat /proc/uptime # showing up time in seconds (left) and idle time (right) (idle time > up time suggests more than one cpu)
cat /proc/loadavg
# to view uptime over a period
watch -n 4 uptime
tload # shows load over one, five, and fifteen minute averages
top and vmstat
top # interactive mode
top -b -n1 # batch mode , 1 iteration
vmstat
vmstat -S k # show in Kilobytes
vmstat -S m # show in Megabytes
vmstat 5 3 # show 3 iterations with a 5 second interval
Using Sysstat to Monitor Performance
Installing sysstat
# as root
yum install -y sysstat
cat /etc/cron.d/sysstat
cat /etc/sysconfig/sysstat
systemctl start sysstat
systemctl enable sysstat
iostat -m
pidstat -p $$ 5 3 #
mpstat -P ALL 2 3 # cpu status
Reporting using sar (System Activity Reporter)
sar -u # CPU
sar -r # Memory
sar -n DEV # Network
sar -q # load averages
sar -s 14:00:00 -e 15:00:00 # between start and end times
ls -l /var/log/sa/ # file sa28 found, corresponds to today's date 28th day of the month
sar -f /var/log/sa/sa28 # for a particular file, i.e. day