Linux Admin - Process Management
Following are the common commands used with Process Management–bg, fg, nohup, ps, pstree, top, kill, killall, free, uptime, nice.
Work with Processes
Quick Note: Process PID in Linux
In Linux every running process is given a PID or Process ID Number. This PID is how CentOS identifies a particular process. As we have discussed, systemd is the first process started and given a PID of 1 in CentOS.
Pgrep is used to get Linux PID for a given process name.
[root@CentOS]# pgrep systemd
1
[root@CentOS]#
As seen, the pgrep command returns the current PID of systemd.
Basic CentOS Process and Job Management in CentOS
When working with processes in Linux it is important to know how basic foregrounding and backgrounding processes is performed at the command line.
fg − Bringsthe process to the foreground
bg − Movesthe process to the background
jobs − List of the current processes attached to the shell
ctrl+z − Control + z key combination to sleep the current process
& − Startsthe process in the background
Let's start using the shell command sleep. sleep will simply do as it is named, sleep for a defined period of time: sleep.
[root@CentOS ~]$ jobs
[root@CentOS ~]$ sleep 10 &
[1] 12454
[root@CentOS ~]$ sleep 20 &
[2] 12479
[root@CentOS ~]$ jobs
[1]- Running sleep 10 &
[2]+ Running sleep 20 &
[cnetos@CentOS ~]$
Now, let's bring the first job to the foreground −
[root@CentOS ~]$ fg 1
sleep 10
If you are following along, you'll notice the foreground job is stuck in your shell. Now, let's put the process to sleep, then re-enable it in the background.
- Hit control+z
- Type: bg 1, sending the first job into the background and starting it.
[root@CentOS ~]$ fg 1
sleep 20
^Z
[1]+ Stopped sleep 20
[root@CentOS ~]$ bg 1
[1]+ sleep 20 &
[root@CentOS ~]$
nohup
When working from a shell or terminal, it is worth noting that by default all the processes and jobs attached to the shell will terminate when the shell is closed or the user logs out. When using nohup the process will continue to run if the user logs out or closes the shell to which the process is attached.
[root@CentOS]# nohup ping www.google.com &
[1] 27299
nohup: ignoring input and appending output to ‘nohup.out’
[root@CentOS]# pgrep ping
27299
[root@CentOS]# kill -KILL `pgrep ping`
[1]+ Killed nohup ping www.google.com
[root@CentOS rdc]# cat nohup.out
PING www.google.com (216.58.193.68) 56(84) bytes of data.
64 bytes from sea15s07-in-f4.1e100.net (216.58.193.68): icmp_seq = 1 ttl = 128
time = 51.6 ms
64 bytes from sea15s07-in-f4.1e100.net (216.58.193.68): icmp_seq = 2 ttl = 128
time = 54.2 ms
64 bytes from sea15s07-in-f4.1e100.net (216.58.193.68): icmp_seq = 3 ttl = 128
time = 52.7 ms
ps Command
The ps command is commonly used by administrators to investigate snapshots of a specific process. ps is commonly used with grep to filter out a specific process to analyze.
[root@CentOS ~]$ ps axw | grep python
762 ? Ssl 0:01 /usr/bin/python -Es /usr/sbin/firewalld --nofork -nopid
1296 ? Ssl 0:00 /usr/bin/python -Es /usr/sbin/tuned -l -P
15550 pts/0 S+ 0:00 grep --color=auto python
In the above command, we see all the processes using the python interpreter. Also included with the results were our grep command, looking for the string python.
Following are the most common command line switches used with ps.
| Switch |
Action |
| a |
Excludes constraints of only the reporting processes for the current user |
| x |
Shows processes not attached to a tty or shell |
| w |
Formats wide output display of the output |
| e |
Shows environment after the command |
| -e |
Selects all processes |
| -o |
User-defined formatted output |
| -u |
Shows all processes by a specific user |
| -C |
Shows all processes by name or process id |
| --sort |
Sorts the processes by definition |
To see all processes in use by the nobody user −
[root@CentOS ~]$ ps -u nobody
PID TTY TIME CMD
1853 ? 00:00:00 dnsmasq
[root@CentOS ~]$
To see all information about the firewalld process −
[root@CentOS ~]$ ps -wl -C firewalld
F S UID PID PPID C PRI NI ADDR SZ WCHAN TTY TIME CMD
0 S 0 762 1 0 80 0 - 81786 poll_s ? 00:00:01 firewalld
[root@CentOS ~]$
Let's see which processes are consuming the most memory −
[root@CentOS ~]$ ps aux --sort=-pmem | head -10
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
cnetos 6130 0.7 5.7 1344512 108364 ? Sl 02:16 0:29 /usr/bin/gnome-shell
cnetos 6449 0.0 3.4 1375872 64440 ? Sl 02:16 0:00 /usr/libexec/evolution-calendar-factory
root 5404 0.6 2.1 190256 39920 tty1 Ssl+ 02:15 0:27 /usr/bin/Xorg :0 -background none -noreset -audit 4 -verbose -auth /run/gdm/auth-for-gdm-iDefCt/database -seat seat0 -nolisten tcp vt1
cnetos 6296 0.0 1.7 1081944 32136 ? Sl 02:16 0:00 /usr/libexec/evolution/3.12/evolution-alarm-notify
cnetos 6350 0.0 1.5 560728 29844 ? Sl 02:16 0:01 /usr/bin/prlsga
cnetos 6158 0.0 1.4 1026956 28004 ? Sl 02:16 0:00 /usr/libexec/gnome-shell-calendar-server
cnetos 6169 0.0 1.4 1120028 27576 ? Sl 02:16 0:00 /usr/libexec/evolution-source-registry
root 762 0.0 1.4 327144 26724 ? Ssl 02:09 0:01 /usr/bin/python -Es /usr/sbin/firewalld --nofork --nopid
cnetos 6026 0.0 1.4 1090832 26376 ? Sl 02:16 0:00 /usr/libexec/gnome-settings-daemon
[root@CentOS ~]$
See all the processes by user centos and format, displaying the custom output −
[cnetos@CentOS ~]$ ps -u cnetos -o pid,uname,comm
PID USER COMMAND
5802 centos gnome-keyring-d
5812 cnetos gnome-session
5819 cnetos dbus-launch
5820 cnetos dbus-daemon
5888 cnetos gvfsd
5893 cnetos gvfsd-fuse
5980 cnetos ssh-agent
5996 cnetos at-spi-bus-laun
pstree Command
pstree is similar to ps but is not often used. It displays the processes in a neater tree fashion.
[centos@CentOS ~]$ pstree
systemd─┬─ModemManager───2*[{ModemManager}]
├─NetworkManager─┬─dhclient
│ └─2*[{NetworkManager}]
├─2*[abrt-watch-log]
├─abrtd
├─accounts-daemon───2*[{accounts-daemon}]
├─alsactl
├─at-spi-bus-laun─┬─dbus-daemon───{dbus-daemon}
│ └─3*[{at-spi-bus-laun}]
├─at-spi2-registr───2*[{at-spi2-registr}]
├─atd
├─auditd─┬─audispd─┬─sedispatch
│ │ └─{audispd}
│ └─{auditd}
├─avahi-daemon───avahi-daemon
├─caribou───2*[{caribou}]
├─cgrulesengd
├─chronyd
├─colord───2*[{colord}]
├─crond
├─cupsd
The total output from pstree can exceed 100 lines. Usually, ps will give more useful information.
top Command
top is one of the most often used commands when troubleshooting performance issues in Linux. It is useful for real-time stats and process monitoring in Linux. Following is the default output of top when brought up from the command line.
Tasks: 170 total, 1 running, 169 sleeping, 0 stopped, 0 zombie
%Cpu(s): 2.3 us, 2.0 sy, 0.0 ni, 95.7 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 1879668 total, 177020 free, 607544 used, 1095104 buff/cache
KiB Swap: 3145724 total, 3145428 free, 296 used. 1034648 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5404 root 20 0 197832 48024 6744 S 1.3 2.6 1:13.22 Xorg
8013 centos 20 0 555316 23104 13140 S 1.0 1.2 0:14.89 gnome-terminal-
6339 centos 20 0 332336 6016 3248 S 0.3 0.3 0:23.71 prlcc
6351 centos 20 0 21044 1532 1292 S 0.3 0.1 0:02.66 prlshprof
Common hot keys used while running top (hot keys are accessed by pressing the key as top is running in your shell).
| Command |
Action |
| b |
Enables / disables bold highlighting on top menu |
| z |
Cycles the color scheme |
| l |
Cycles the load average heading |
| m |
Cycles the memory average heading |
| t |
Task information heading |
| h |
Help menu |
| Shift+F |
Customizes sorting and display fields |
Following are the common command line switches for top.
| Command |
Action |
| -o |
Sorts by column (can prepend with - or + to sort ascending or descending) |
| -u |
Shows only processes from a specified user |
| -d |
Updates the delay time of top |
| -O |
Returns a list of columns which top can apply sorting |
Sorting options screen in top, presented using Shift+F. This screen allows customization of top display and sort options.
Fields Management for window 1:Def, whose current sort field is %MEM
Navigate with Up/Dn, Right selects for move then <Enter> or Left commits,
'd' or <Space> toggles display, 's' sets sort. Use 'q' or <Esc> to end!
* PID = Process Id TGID = Thread Group Id
* USER = Effective User Name ENVIRON = Environment vars
* PR = Priority vMj = Major Faults delta
* NI = Nice Value vMn = Minor Faults delta
* VIRT = Virtual Image (KiB) USED = Res+Swap Size (KiB)
* RES = Resident Size (KiB) nsIPC = IPC namespace Inode
* SHR = Shared Memory (KiB) nsMNT = MNT namespace Inode
* S = Process Status nsNET = NET namespace Inode
* %CPU = CPU Usage nsPID = PID namespace Inode
* %MEM = Memory Usage (RES) nsUSER = USER namespace Inode
* TIME+ = CPU Time, hundredths nsUTS = UTS namespace Inode
* COMMAND = Command Name/Line
PPID = Parent Process pid
UID = Effective User Id
top, showing the processes for user rdc and sorted by memory usage −
PID USER %MEM PR NI VIRT RES SHR S %CPU TIME+ COMMAND
6130 rdc 6.2 20 0 1349592 117160 33232 S 0.0 1:09.34 gnome-shell
6449 rdc 3.4 20 0 1375872 64428 21400 S 0.0 0:00.43 evolution-calen
6296 rdc 1.7 20 0 1081944 32140 22596 S 0.0 0:00.40 evolution-alarm
6350 rdc 1.6 20 0 560728 29844 4256 S 0.0 0:10.16 prlsga
6281 rdc 1.5 20 0 1027176 28808 17680 S 0.0 0:00.78 nautilus
6158 rdc 1.5 20 0 1026956 28004 19072 S 0.0 0:00.20 gnome-shell-cal
Showing valid top fields (condensed) −
[centos@CentOS ~]$ top -O
PID
PPID
UID
USER
RUID
RUSER
SUID
SUSER
GID
GROUP
PGRP
TTY
TPGID
kill Command
The kill command is used to kill a process from the command shell via its PID. When killing a process, we need to specify a signal to send. The signal lets the kernel know how we want to end the process. The most commonly used signals are −
SIGTERM is implied as the kernel lets a process know it should stop soon as it is safe to do so. SIGTERM gives the process an opportunity to exit gracefully and perform safe exit operations.
SIGHUP most daemons will restart when sent SIGHUP. This is often used on the processes when changes have been made to a configuration file.
SIGKILL since SIGTERM is the equivalent to asking a process to shut down. The kernel needs an option to end a process that will not comply with requests. When a process is hung, the SIGKILL option is used to shut the process down explicitly.
For a list off all signals that can be sent with kill the -l option can be used −
[root@CentOS]# kill -l
1) SIGHUP 2) SIGINT 3) SIGQUIT 4) SIGILL 5) SIGTRAP
6) SIGABRT 7) SIGBUS 8) SIGFPE 9) SIGKILL 10) SIGUSR1
11) SIGSEGV 12) SIGUSR2 13) SIGPIPE 14) SIGALRM 15) SIGTERM
16) SIGSTKFLT 17) SIGCHLD 18) SIGCONT 19) SIGSTOP 20) SIGTSTP
21) SIGTTIN 22) SIGTTOU 23) SIGURG 24) SIGXCPU 25) SIGXFSZ
26) SIGVTALRM 27) SIGPROF 28) SIGWINCH 29) SIGIO 30) SIGPWR
31) SIGSYS 34) SIGRTMIN 35) SIGRTMIN+1 36) SIGRTMIN+2 37) SIGRTMIN+3
38) SIGRTMIN+4 39) SIGRTMIN+5 40) SIGRTMIN+6 41) SIGRTMIN+7 42) SIGRTMIN+8
43) SIGRTMIN+9 44) SIGRTMIN+10 45) SIGRTMIN+11 46) SIGRTMIN+12 47) SIGRTMIN+13
48) SIGRTMIN+14 49) SIGRTMIN+15 50) SIGRTMAX-14 51) SIGRTMAX-13 52) SIGRTMAX-12
53) SIGRTMAX-11 54) SIGRTMAX-10 55) SIGRTMAX-9 56) SIGRTMAX-8 57) SIGRTMAX-7
58) SIGRTMAX-6 59) SIGRTMAX-5 60) SIGRTMAX-4 61) SIGRTMAX-3 62) SIGRTMAX-2
63) SIGRTMAX-1 64) SIGRTMAX
[root@CentOS rdc]#
Using SIGHUP to restart system.
[root@CentOS]# pgrep systemd
1
464
500
643
15071
[root@CentOS]# kill -HUP 1
[root@CentOS]# pgrep systemd
1
464
500
643
15196
15197
15198
[root@CentOS]#
pkill will allow the administrator to send a kill signal by the process name.
[root@CentOS]# pgrep ping
19450
[root@CentOS]# pkill -9 ping
[root@CentOS]# pgrep ping
[root@CentOS]#
killall will kill all the processes. Be careful using killall as root, as it will kill all the processes for all users.
[root@CentOS]# killall chrome
free Command
free is a pretty simple command often used to quickly check the memory of a system. It displays the total amount of used physical and swap memory.
[root@CentOS]# free
total used free shared buff/cache available
Mem: 1879668 526284 699796 10304 653588 1141412
Swap: 3145724 0 3145724
[root@CentOS]#
nice Command
nice will allow an administrator to set the scheduling priority of a process in terms of CPU usages. The niceness is basically how the kernel will schedule CPU time slices for a process or job. By default, it is assumed the process is given equal access to CPU resources.
First, let's use top to check the niceness of the currently running processes.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
28 root 39 19 0 0 0 S 0.0 0.0 0:00.17 khugepaged
690 root 39 19 16808 1396 1164 S 0.0 0.1 0:00.01 alsactl]
9598 rdc 39 19 980596 21904 10284 S 0.0 1.2 0:00.27 tracker-extract
9599 rdc 39 19 469876 9608 6980 S 0.0 0.5 0:00.04 tracker-miner-a
9609 rdc 39 19 636528 13172 8044 S 0.0 0.7 0:00.12 tracker-miner-f
9611 rdc 39 19 469620 8984 6496 S 0.0 0.5 0:00.02 tracker-miner-u
27 root 25 5 0 0 0 S 0.0 0.0 0:00.00 ksmd
637 rtkit 21 1 164648 1276 1068 S 0.0 0.1 0:00.11 rtkit-daemon
1 root 20 0 128096 6712 3964 S 0.3 0.4 0:03.57 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:00.01 kthreadd
3 root 20 0 0 0 0 S 0.0 0.0 0:00.50 ksoftirqd/0
7 root 20 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
8 root 20 0 0 0 0 S 0.0 0.0 0:00.00 rcu_bh
9 root 20 0 0 0 0 S 0.0 0.0 0:02.07 rcu_sched
We want to focus on the NICE column depicted by NI. The niceness range can be anywhere between -20 to positive 19. -20 represents the highest given priority.
nohup nice --20 ping www.google.com &
renice
renice allows us to change the current priority of a process that is already running.
renice 17 -p 30727
The above command will lower the priority of our ping process command.
