Opened 8 years ago

Last modified 2 years ago

#737 new defect

Running away with memory

Reported by: alexcescalona@… Owned by: quinox
Priority: normal Milestone: Release 1.3.0
Component: nicotine Version: 1.2.16
Keywords: Cc: alexcescalona@…

Description

After a day or two of leaving Nicotine+ to its own devices, it beginnings to run away with memory, consuming as much as is available. At that point, I kill the process and then relaunch it.

Some information about my user environment. I'm running Ubuntu 10.04 (32-bit), 2.6.32-28-generic-pae. Also, Nicotine+ often runs concurrently in two separate user accounts on the same system.

I've not looked into limiting the amount of memory available to it...

Change History (3)

comment:1 Changed 8 years ago by quinox

I have debugged leaking memory a few months ago, I could not recreate it: during my tests all objects were released as required. Since python manages the memory releasing objects doesn't directly cause a drop in memory usage, so if you have a lot of free memory it's possible python keeps "using" more of it. This is probably virtual memory which doesn't count, you can have a process use 100GB of virtual memory without using a single byte in the RAM.

Please do try to limit N+. If it starts crashing it probably is a memory leak somewhere.

Also, Nicotine+ often runs concurrently in two separate user accounts on the same system.

That's impossible for N+ to achieve by itself since it has no root access, so that means you either run it twice on different accounts or it's a mistake (either it's not N+ or the information shown isn't correct)

comment:2 Changed 3 years ago by gfarmerfr

Milestone: Release 1.2.17Release 1.3.0

comment:3 Changed 2 years ago by gfarmerfr

From what I gather the memory leak might be located in the "respond to distributed searches" functionality. I need to investigate more.

Note: See TracTickets for help on using tickets.