Super User is a question and answer site for computer enthusiasts and power users. It only takes a minute to sign up.
Sign up to join this community
I like to use htop to show me how the server is being used.
Unfortunately with modern servers, the machine might have 48 or even 120 cores. This means that I can only see the first few lines of htop and never the second half of the htop display that shows which processes are using the machine.
How can I hide all of the core-usage lines or even better aggregate them for some sort of statistic?
Open the setup screen using F2 or ShiftS. The first page of that screen is dedicated to configuring header meters, so you can remove "CPUs" and add "CPU average".
In recent htop versions, there also are "CPUs (1&2/4)" and "CPUs (3&4/4)" meters showing two cores per line, though this is more of use on 8–32 core systems.
To save two more lines, open the "Display options" page and turn off "Leave a margin around header".
F2 in ssh. You can use F2 to edit htop in your local machine. And diff the config file(~/.config/htop/htoprc) before and after change. Then apply the diff to your remote machine.
Based on grawity's answer, you can create a configuration that you like on a different machine and then copy it to the machine where the problem occurs.
The configuration is saved (under Debian) under ~/.config/htop/htoprc.
F2 to get into the configuration.~/.config/htop/htoprc to the larger machine.In my case (120 cores) a configuration with "CPUs (1&2/4)" on the left side of the header, and "CPUs (3&4/4)" on the right side of the header looks good. As a result, the header takes up about half the screen and the other half lists the processes. Each line in the header shows four CPUs which is fine for me.
Sample configuration:
# Beware! This file is rewritten by htop when settings are changed in the interface.
# The parser is also very primitive, and not human-friendly.
fields=0 48 17 18 38 39 40 2 46 47 49 1
sort_key=46
sort_direction=1
hide_threads=0
hide_kernel_threads=1
hide_userland_threads=0
shadow_other_users=0
show_thread_names=0
highlight_base_name=0
highlight_megabytes=1
highlight_threads=0
tree_view=0
header_margin=1
detailed_cpu_time=0
cpu_count_from_zero=0
color_scheme=0
delay=15
left_meters=Memory Swap CPU Load LoadAverage
left_meter_modes=1 1 1 1 1
right_meters=Tasks LoadAverage Uptime
right_meter_modes=2 2 2
#Alternative (Blind navigation)
Press F2, left, F9. (If CPUs are the items in the header.) After this you can see what is going on and would continue by pressing F10 to quit the configuration.
Thanks to islandman93:
New blind navigation:
F2,right,delete,right,delete. Then you probably want to add cpu average to the left column
F2, right, delete, right, delete. Then you probably want to add cpu average to the left column
Jul 23, 2020 at 15:23
right_meters, right_meters and corresponding left_meter_modes, right_meter_modes.
Try the 't' key.
None of the other answers helped. My top and terminal must be different. My top was installed via the procps-3.2.8-45.0.1.el6_9.1.x86_64 package on Oracle Enterprise Linux (repackaged RedHat Enterprise Linux) 6.9 and I was accessing it via PuTTY 0.62.
I just had this issue as well, system has 24 cores, boatloads of disks and interfaces, and I couldn't read the process data after all the mem/disk/net lines etc..
Simply starting it differently was the easiest solution:
atop -l
From the man page: Limit the number of system level lines for the counters per-cpu, the active disks and the network interfaces.
SuSE : Press F2, Press F10, press q, sed -i 's/AllCPUs/CPU/g' ~/.htoprc
debian : Press F2, Press F10, press q, sed -i 's/AllCPUs/CPU/g' ~/.config/htop/htoprc
