Insame amount of used RAM, unable to find the cause

[AlexZap]

Hi, i'm having some prolem with the RAM usage, after boot i nearly have 2G of used ram but after two days it reach 8G, in about a week it fill 24G of ram, and at that point i need to reboot the system.

doing the command ps aux | awk '{print $6/1024 " MB\t\t" $11}' | sort -n i have found an  strange amount of dhcpcd processram.txt.

 

graph of memory usage of the last two days: 

 

casaz-nas-diagnostics-20230819-2301.zip

Edited August 19, 2023 by AlexZap

[JorgeB]

13 hours ago, AlexZap said:

in about a week it fill 24G of ram

That's not used RAM:

https://www.linuxatemyram.com/

[AlexZap]

ok, thank you,  but when the used ram(not the cached) reach 100% the system become unresponsive and every action on the webpage take a very long time to load.

[JorgeB]

Then there's likely some app with a memory leak, post the diagnostics when that happens to see if there's something visible.

 

[AlexZap]

17 hours ago, JorgeB said:

Then there's likely some app with a memory leak, post the diagnostics when that happens to see if there's something visible.

 

now it reached 16G, and it still going up pretty fast. Now there are 3500 more processesram2.txt

casaz-nas-diagnostics-20230820-2159.zip

Edited August 21, 2023 by AlexZap

[AlexZap]

 

18 hours ago, JorgeB said:

Then there's likely some app with a memory leak, post the diagnostics when that happens to see if there's something visible.

 

found the cause, i configured unraid to use only ipv4 and ram usage returned to normal but now i can't re enable ipv6+ipv4, ipv6 memory leak problem?. i think something is wrong with unraid and ipv6, after applying settings nothing change, only reboot apply settings and the server is unable to get a dns server from dhcp

Edited August 21, 2023 by AlexZap

[JorgeB]

Try updating to v6.12.4 once it's out, there are some ipv6 fixes.

[ApfelBirneKreis]

same problem here even with 6.12.4

[Bremsklotz]

I had the same problem. After a couple days my ram became full and I had to restart the server. Otherwise the system was unresponsive.

 

I used Netdata to find the cause: there was something like memory-dhcp which used a lot of ram.

I think it was a network misconfig. I solved it like this:

 

settings --> network --> ip adress assign for ipv4 and ipv6 from automatic to static

settings --> docker --> advanced view --> deactivate custom ipv4 and ipv6 network on br0

 

No problems since I changed that.