ipfs / ipfs-desktop

An unobtrusive and user-friendly desktop application for IPFS on Windows, Mac and Linux.
https://docs.ipfs.tech/install/ipfs-desktop/
MIT License
5.85k stars 849 forks source link

Version 0.37.0 consumes a lot of RAM on Ubuntu-based Linux OS #2833

Open heEXDe opened 1 week ago

heEXDe commented 1 week ago

OS: Ubuntu-based distribution of Linux Version of IPFS Desktop: 0.37.0 (AppImage)

Description:

Version 0.37.0 consumes a lot of RAM (sometimes over 5 GB), when connecting with more than 2K peers.

To reproduce:

  1. start the app
  2. open the system monitor
welcome[bot] commented 1 week ago

Thank you for submitting your first issue to this repository! A maintainer will be here shortly to triage and review. In the meantime, please double-check that you have provided all the necessary information to make this process easy! Any information that can help save additional round trips is useful! We are triaging issues on weekly basis and aim to give initial feedback within a few business days. If this does not happen, feel free to leave a comment. Please keep an eye on how this issue will be labeled, as labels give an overview of priorities, assignments and additional actions requested by the maintainers:

Finally, remember to use https://discuss.ipfs.tech if you just need general support.

lidel commented 3 days ago

@heEXDe thank you for filling this. Mind providing bit more context, so we can route this to the right repository?

  1. is the memory usage coming mostly from kubo daemon or Electron (Desktop app GUI)?
  2. how long is your node running?
  3. are you running with default Kubo configuration, or did you adjust things like Swarm.ConnMgr? The default should trim unused connections down to <100 when idle.
heEXDe commented 3 days ago
  1. Most of the memory usage seems to come from Kubo itself (screenshot) Screenshot_2024-06-27_23-25-57

  2. The node usually operates more than 5 hours a day

  3. I don't recall making changes to the default settings. When idle (connections to fewer than 100 peers), the node doesn't use much less memory (the decrease is about 10%), that's part of the problem.