You are not logged in.
System monitor and top show an increasing amount of ram and CPU usage but no processes that are using that much. The amount of ram being used increases until rebooting. Is there any way to tell what is using up the system recources besides top and System Monitor? Sorry if this is the wrong place for this.
Offline
Likely to be a case of disk caching - see http://www.linuxatemyram.com/
check the output of free, whatever is in the buffers/cache line is RAM that is used for disk caching but can be immediately given to processes when necessary.
Offline
I am fairly confident that it is not disk cacheing as there is a noticable slow down/freeze whenever it occurs. I will have to check free -m when it occurs to be sure.
Last edited by aDragon (2016-04-20 01:12:42)
Offline
Maybe tmpfs eat your RAM.
Offline
Is there any way to tell what is using up the system recources besides top and System Monitor?
`ps_mem` was added to [Community] recently, try that.
Offline
atop is a decent system resources monitor.
I also use this to show the top five memory hungry processes:
ps aux --sort=-%mem | awk 'NR<=5{print $0}'
Another thing to try is watch for processes:
watch ps aux --sort=-%mem
Edited: Typo
Last edited by triforce (2016-04-20 14:37:53)
Offline
I am fairly confident that it is not disk cacheing as there is a noticable slow down/freeze whenever it occurs. I will have to check free -m when it occurs to be sure.
Definetely not cacheing. System monitor counts cached RAM as free.
Offline
"free -m" is showing 1200 used with 6100 buffer/cached. However avalible is only showing 700. I had thought avalible was buff/cache + free but aparently I am wrong.
total used free shared buff/cache available
Mem: 7451 1218 61 3 6171 776
Swap: 7558 116 7442
And "ps aux --sort=-%mem | awk 'NR<=5{print $0}'" isn't showing anything using that much memory.
ps aux --sort=-%mem | awk 'NR<=5{print $0}'
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
sam 1664 54.8 5.3 1816728 409492 tty2 Sl+ 16:29 32:42 firefox
root 27902 86.4 1.7 182100 130076 ? Ds 17:26 2:07 /usr/lib/systemd/systemd-journald
sam 809 8.5 1.6 1956896 127116 tty2 Sl+ 16:28 5:09 /usr/bin/gnome-shell
sam 26962 4.6 1.3 791872 99652 tty2 Sl+ 17:23 0:13 cool-retro-term
Last edited by aDragon (2016-04-26 21:32:15)
Offline
Yes buff cache is disk cache, very possible that used tmpfs space is counted as such as well. But as mentioned earlier, if you start to start more applications this can be immediately passed on. Are you doing something with a lot of writes? (Well firefox usually does quite a few which will fill up the disk cache)
Also:
man free
available is an estimation of memory available during start of new applications without starting to swap.
You might also want to check the output of df to see if your tmpfs is abnormally full, it should usually be very small usage wise.
Last edited by V1del (2016-04-26 22:46:48)
Offline
tmps doesn't seem to be very high according to df
Filesystem 1K-blocks Used Available Use% Mounted on
dev 3812376 0 3812376 0% /dev
run 3815044 776 3814268 1% /run
/dev/sda8 455205256 289188416 142870652 67% /
tmpfs 3815044 12868 3802176 1% /dev/shm
tmpfs 3815044 0 3815044 0% /sys/fs/cgroup
tmpfs 3815044 64 3814980 1% /tmp
/dev/sda2 262144 99188 162956 38% /boot
tmpfs 763012 16 762996 1% /run/user/120
tmpfs 763012 52 762960 1% /run/user/1000
I tested not running Firefox and while that seemed to help, it still eventually occurred after a while. So, I am not sure if it was just in my head or not.
However, system journal and kworker seem to be using an fairly large amount of CPU.
ps aux --sort=-%cpu | awk 'NR<=5{print $0}'
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 22616 77.5 1.6 173928 123960 ? Rs 23:24 0:22 /usr/lib/systemd/systemd-journald
root 92 57.6 0.0 0 0 ? S 22:33 29:51 [kworker/0:2]
root 38 32.3 0.0 0 0 ? R 22:33 16:45 [kworker/0:1]
Offline