Out-of-memory wrong setting?

Hi

i recognized out-of-memory messages in dmesg, and guess the memory-block in dts is wrong

actual value is afair changed by a patch from lede (in mainline same value is used but in 4.14 it is 0x40000000)

https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts?h=v4.14.64#n102

@moore @linkerosa @Ryder.Lee

can anybody explain me the values? how is size of memory calculated

regards Frank

base = 0x80000000 size = 0x80000000 for (2GB RAM)

1 Like

So it should be correct…any suggestion to prevent oom-messages? I tried vm-settings in /etc/sysctl.conf

# Prevent OOM killer (Out of Memory) which causes the server to crash
vm.overcommit_memory = 2
vm.overcommit_ratio = 80

It may caused by … 1.try to run too many features in run time 2.app/driver has memory leak issue

I will suggest to dump free memory regularly after system boot up to see the trend of free memory, thanks.

System should swap if there is too few memory.

I have this issue after near 15 days (in dmesg)

And aditional to router function i have 2 lxc vms running (webserver+mysql,vncserver+firefox)

Which command do you suggest to get (per-app) memory usage (ps?)?

linux kernel will try to kill applications to free more memory to satisfy the need. you can use ps or top command to show the memory usage, thanks.

ps -eo pid,ppid,cmd,%mem,%cpu --sort=-%mem | head

top -b -o +%MEM | head -n 20