Opened 5 years ago
Last modified 4 years ago
#3423 pending bug
Memory Leak on Linux
Reported by: | Cameron Tacklind | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | not applicable |
Component: | libtorrent | Version: | 2.0.3 |
Keywords: | Cc: |
Description
I've been running Deluged 2.0.3 on my server for a while now. I have nearly 500 torrents seeding but only a few active at a time. It's mostly stable, except every ~month I take a look and it's using 100% of available memory (32GB) and CPU.
Restarting the service brings the memory usage back down to ~400MB. However it starts growing pretty quickly. I restarted my current instance a day or two ago and it's currently using nearly 5GB. I just did a quick test and looked at the RSS memory usage across a 10 minute interval. It seems to be growing at 182 kBps. (At this rate it should only take 2 days to fill all 32GB! So clearly this 10min sample doesn't capture the whole picture...)
I have a nearly identical instance of Deluged running on the same machine for a separate network. It only runs ~50 torrents and never seems to have these memory leak problems.
Change History (5)
comment:1 by , 5 years ago
comment:2 by , 5 years ago
This is the libtorrent issue which was finally fixed in 1.2.10: https://github.com/qbittorrent/qBittorrent/issues/12326
comment:3 by , 5 years ago
Glad to hear. What's it going to take for this to be incorporated to Deluge? Should I be able to update to latest libtorrent, reinstall deluge python3 bindings, and have it work?
comment:4 by , 5 years ago
Yes, just compile libtorrent with python bindings using the latest commit from the 1.2(not 2.0!) branch(if your distro doesn't already provide it in the official repos), and it should solve your problem. I had been experiencing exactly the same issue that you describe, and 1.2.10 completely solved it for me.
comment:5 by , 4 years ago
Component: | Core → libtorrent |
---|---|
Milestone: | needs verified → not applicable |
Status: | new → pending |
I think this is actually a duplicate of #3327. I found some logs spewing out many "Unhandled error in Deferred"s that match what is described there.