Hi, so i got this huge ram leak on manjaro and dont know what to do about,
htop doesn’t listed any proccess that could be using this amount on ram
i have been calculated that is about 4.5 GB
i have removed discord and wine since some pages listed this apps as possibles causes but nothing, i even have upgraded the kernel form 5.9 to 5.11 but the problem stay the same.
pd: sorry for the bad english
Hi and welcome to the forum 
For people here to help you better more data is needed. Check this post on how to provide it.
Readers probably want to know:
Since it is beneficial to have used code in RAM what do you see on your system that is a RAM leak in your case?
How did you determine that this is a leak and can you share some more info?
Removing applications such as wine or discord will not lower RAM use when not started.
For example the output of
free --human --total
would give a picture of your RAM in use on your system.
~]$ free --human --total
total used free shared buff/cache available
Mem: 15Gi 9,6Gi 572Mi 233Mi 5,3Gi 5,4Gi
Swap: 29Gi 1,0Mi 29Gi
Total: 45Gi 9,6Gi 30Gi
thanks for the help, as can be seen here, there is 15 gb aprox installed on the system, 5.3 gb used and 5.4 available, so where is the missing 4 gb? thats is a ram leak?
maybe it shouldn’t be called a leak because the 4 gb missing are permanent.
Thanks for the output, this seems like a normal system. Memory info is not so straight forward. I’m not good at explaining but there is a app site for that: https://www.linuxatemyram.com/ . I hope this explains it for you.
Thanks for the info
free --human --total
total used free shared buff/cache available
Mem: 15Gi 4,7Gi 10Gi 80Mi 684Mi 10Gi
Swap: 29Gi 0B 29Gi
Total: 45Gi 4,7Gi 40Gi
so this is a capture on a fresh restart, there are 4.7gb used, what i had understood is that this 4 gb should be listed as buff/cache, so the given page doesn’t explain why there are 4 gb used in a fresh restart since this is not “buff/cache”. i have been thinking that maybe this memory is reserved for the integrated gpu the intel hd 630 uhd, but i cannot say for sure.
You need to remember that /tmp is mounted on tmpfs which is ram based - so nothing is missing.
Also for your amusement read the linux-ate-my-ram link
i don’t get it, can you give me more info about the relation betwen tmp, tmpfs and the ram? i have been reading about but it is supposed that this files should be cleaned after a restart, but there arent.
But in linuxatemyram.com-play.html i manage to run the test and it run until 10814 MB, so the there are 4 gb that i can’t use and are not listed in htop process
after a wandering the web searching for clues in the matter i find the file /proc/meminfo, so i process it in calc i get the following table, i think that it must explain where is my 4 gb but i dont know what means anything below “swapCached”. some halp pls 
( i get this info whitout aplications running .)
| MemTotal: | 16265712 | kB | 15,884 | mb | 15.512 | gb |
|---|---|---|---|---|---|---|
| MemFree: | 9791528 | kB | 9,562 | mb | 9.338 | gb |
| MemAvailable: | 10221992 | kB | 9,982 | mb | 9.748 | gb |
| Buffers: | 41356 | kB | 40 | mb | 0.039 | gb |
| Cached: | 871020 | kB | 851 | mb | 0.831 | gb |
| SwapCached: | 0 | kB | 0 | mb | 0.000 | gb |
| Active: | 285484 | kB | 279 | mb | 0.272 | gb |
| Inactive: | 1568904 | kB | 1,532 | mb | 1.496 | gb |
| Active(anon): | 1560 | kB | 2 | mb | 0.001 | gb |
| Inactive(anon): | 1184752 | kB | 1,157 | mb | 1.130 | gb |
| Active(file): | 283924 | kB | 277 | mb | 0.271 | gb |
| Inactive(file): | 384152 | kB | 375 | mb | 0.366 | gb |
| Unevictable: | 115604 | kB | 113 | mb | 0.110 | gb |
| Mlocked: | 64 | kB | 0 | mb | 0.000 | gb |
| SwapTotal: | 0 | kB | 0 | mb | 0.000 | gb |
| SwapFree: | 0 | kB | 0 | mb | 0.000 | gb |
| Dirty: | 728 | kB | 1 | mb | 0.001 | gb |
| Writeback: | 0 | kB | 0 | mb | 0.000 | gb |
| AnonPages: | 1058036 | kB | 1,033 | mb | 1.009 | gb |
| Mapped: | 410836 | kB | 401 | mb | 0.392 | gb |
| Shmem: | 244020 | kB | 238 | mb | 0.233 | gb |
| KReclaimable: | 37956 | kB | 37 | mb | 0.036 | gb |
| Slab: | 113284 | kB | 111 | mb | 0.108 | gb |
| SReclaimable: | 37956 | kB | 37 | mb | 0.036 | gb |
| SUnreclaim: | 75328 | kB | 74 | mb | 0.072 | gb |
| KernelStack: | 10928 | kB | 11 | mb | 0.010 | gb |
| PageTables: | 17812 | kB | 17 | mb | 0.017 | gb |
| NFS_Unstable: | 0 | kB | 0 | mb | 0.000 | gb |
| Bounce: | 0 | kB | 0 | mb | 0.000 | gb |
| WritebackTmp: | 0 | kB | 0 | mb | 0.000 | gb |
| CommitLimit: | 6035704 | kB | 5,894 | mb | 5.756 | gb |
| Committed_AS: | 4098104 | kB | 4,002 | mb | 3.908 | gb |
| VmallocTotal: | 34359738367 | kB | 33,554,432 | mb | 32768.000 | gb |
| VmallocUsed: | 89700 | kB | 88 | mb | 0.086 | gb |
| VmallocChunk: | 0 | kB | 0 | mb | 0.000 | gb |
| Percpu: | 4736 | kB | 5 | mb | 0.005 | gb |
| HardwareCorrupted: | 0 | kB | 0 | mb | 0.000 | gb |
| AnonHugePages: | 0 | kB | 0 | mb | 0.000 | gb |
| ShmemHugePages: | 0 | kB | 0 | mb | 0.000 | gb |
| ShmemPmdMapped: | 0 | kB | 0 | mb | 0.000 | gb |
| FileHugePages: | 0 | kB | 0 | mb | 0.000 | gb |
| FilePmdMapped: | 0 | kB | 0 | mb | 0.000 | gb |
| CmaTotal: | 0 | kB | 0 | mb | 0.000 | gb |
| CmaFree: | 0 | kB | 0 | mb | 0.000 | gb |
| HugePages_Total: | 2048 | 2 | mb | 0.002 | gb | |
| HugePages_Free: | 2048 | 2 | mb | 0.002 | gb | |
| HugePages_Rsvd: | 0 | 0 | mb | 0.000 | gb | |
| HugePages_Surp: | 0 | 0 | mb | 0.000 | gb | |
| Hugepagesize: | 2048 | kB | 2 | mb | 0.002 | gb |
| Hugetlb: | 4194304 | kB | 4,096 | mb | 4.000 | gb |
| DirectMap4k: | 275136 | kB | 269 | mb | 0.262 | gb |
| DirectMap2M: | 6936576 | kB | 6,774 | mb | 6.615 | gb |
| DirectMap1G: | 9437184 | kB | 9,216 | mb | 9.000 | gb |
| 0 | mb | 0.000 | gb |
I am more inclined to think this is hardware related - somehow - I don’t know - but if you system is build with 4x4G modules and one is flaky - that would explain a lot.
So I suggest you start troubleshooting by pulling 3 blocks - and reboot to test.
If that is not possible then boot the system using memtest+ ISO and run some testing
sudo pacman -Syu memtest86-efi
reboot and hold Shift to access grub menu - select memtest
To get a more accurate ram usage by every processes, you can install ps_mem and run it as root.
Eg. in my case:
sudo ps_mem
Private + Shared = RAM used Program
380.0 KiB + 94.5 KiB = 474.5 KiB rtkit-daemon
396.0 KiB + 80.5 KiB = 476.5 KiB incrond
620.0 KiB + 135.5 KiB = 755.5 KiB gvfsd-metadata
884.0 KiB + 110.5 KiB = 994.5 KiB dconf-service
996.0 KiB + 382.5 KiB = 1.3 MiB gvfsd
1.0 MiB + 423.5 KiB = 1.5 MiB gvfsd-trash
1.2 MiB + 743.5 KiB = 2.0 MiB systemd-logind
1.7 MiB + 409.5 KiB = 2.1 MiB thermald
2.0 MiB + 465.5 KiB = 2.4 MiB sudo
2.4 MiB + 191.5 KiB = 2.6 MiB picom
2.9 MiB + 41.5 KiB = 2.9 MiB xbindkeys
2.3 MiB + 701.5 KiB = 2.9 MiB upowerd
2.4 MiB + 579.0 KiB = 3.0 MiB dbus-daemon (2)
3.0 MiB + 111.5 KiB = 3.1 MiB xfconfd
2.7 MiB + 626.5 KiB = 3.3 MiB systemd-udevd
3.2 MiB + 121.5 KiB = 3.4 MiB gvfsd-fuse
3.1 MiB + 624.5 KiB = 3.7 MiB rngd
4.0 MiB + 175.5 KiB = 4.2 MiB accounts-daemon
3.8 MiB + 610.5 KiB = 4.4 MiB wpa_supplicant
3.9 MiB + 796.0 KiB = 4.7 MiB lightdm (2)
4.2 MiB + 801.5 KiB = 5.0 MiB gvfs-udisks2-volume-monitor
6.7 MiB + 697.5 KiB = 7.3 MiB polkit-gnome-authentication-agent-1
6.8 MiB + 780.5 KiB = 7.6 MiB udisksd
6.9 MiB + 706.5 KiB = 7.6 MiB xfce4-volumed-pulse
7.0 MiB + 726.5 KiB = 7.7 MiB xfce4-power-manager
7.1 MiB + 837.5 KiB = 7.9 MiB panel-19-systra
7.4 MiB + 839.5 KiB = 8.2 MiB xfce4-session
7.4 MiB + 775.5 KiB = 8.2 MiB Thunar
7.9 MiB + 875.5 KiB = 8.7 MiB python3.9
7.4 MiB + 1.4 MiB = 8.8 MiB NetworkManager
7.8 MiB + 1.7 MiB = 9.5 MiB xfce4-notifyd
9.6 MiB + 564.5 KiB = 10.2 MiB pulseaudio
9.8 MiB + 900.5 KiB = 10.6 MiB xfsettingsd
10.7 MiB + 833.5 KiB = 11.5 MiB xfce4-screensaver
10.2 MiB + 1.8 MiB = 12.0 MiB panel-15-power-
11.1 MiB + 2.0 MiB = 13.0 MiB xfce4-panel
10.7 MiB + 2.3 MiB = 13.0 MiB nm-applet
8.9 MiB + 4.4 MiB = 13.4 MiB systemd (3)
11.8 MiB + 2.0 MiB = 13.8 MiB conky (3)
13.6 MiB + 791.5 KiB = 14.4 MiB systemd-journald
11.8 MiB + 2.8 MiB = 14.6 MiB xfce4-terminal
14.7 MiB + 74.5 KiB = 14.7 MiB bash
14.7 MiB + 320.5 KiB = 15.0 MiB polkitd
13.1 MiB + 2.2 MiB = 15.3 MiB xfwm4
14.2 MiB + 2.7 MiB = 16.8 MiB panel-4-whisker
14.8 MiB + 2.1 MiB = 17.0 MiB panel-9-docklik
18.9 MiB + 1.1 MiB = 20.0 MiB xfdesktop
25.3 MiB + 384.5 KiB = 25.7 MiB plugin_host
29.4 MiB + 3.2 MiB = 32.6 MiB autokey-gtk
29.6 MiB + 4.8 MiB = 34.4 MiB gedit
36.9 MiB + 2.3 MiB = 39.2 MiB kdeconnectd
33.5 MiB + 9.5 MiB = 43.0 MiB Xorg
83.2 MiB + 16.5 KiB = 83.2 MiB dnscrypt-proxy
260.1 MiB + 8.7 MiB = 268.8 MiB subl3
441.5 MiB + 101.6 MiB = 543.0 MiB firefox (4)
---------------------------------
1.4 GiB
This is the result after a fresh reboot, after 8 minutes when I launched Firefox with many addons; I have 16 GB of RAM.
After some day of uptime, I can also reach 4 and more GB of used RAM, but especially high value of cache from free -h (also 10 GB of cache); the system is always smooth and reactive.
Yep - that is what the author of linux-ate-my-ram is trying to explain - Linux cache.
Yep: I want more used cache as my system can handle; in facts I set
vm.vfs_cache_pressure=5 and the system, as I’ve said here and elsewhere, is amazingly smooth and fast.
I use the same technique + I have enabled zswap too - and my experience is like yours … very good.
so i install and run ps_mem command but the showed info differ from the given by free or htop
$ sudo ps_mem
Private + Shared = RAM used Program
412.0 KiB + 484.0 KiB = 896.0 KiB xcape
...
81.1 MiB + 105.7 MiB = 186.8 MiB Xorg
---------------------------------
1.3 GiB
=================================
$ free
total used free shared buff/cache available
Mem: 15884 4789 10354 85 740 10730
Swap: 30719 0 30719