Skip to content

Update tiny network tweaks 99-cachyos-settings.conf #187

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
SoongVilda wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Update tiny network tweaks 99-cachyos-settings.conf #187

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 31 additions & 5 deletions usr/lib/sysctl.d/99-cachyos-settings.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ vm.page-cluster = 0
# flusher threads will start writing out dirty data.
vm.dirty_background_bytes = 67108864

# The kernel flusher threads will periodically wake up and write old data out to disk. This
# The kernel flusher threads will periodically wake up and write old data out to disk.  This
# tunable expresses the interval between those wakeups, in 100'ths of a second (Default is 500).
vm.dirty_writeback_centisecs = 1500

Expand All @@ -38,9 +38,35 @@ kernel.printk = 3 3 3 3
# Restricting access to kernel pointers in the proc filesystem
kernel.kptr_restrict = 2

# Increase netdev receive queue
# May help prevent losing packets
net.core.netdev_max_backlog = 4096

# Set size of file handles and inode cache
# (Often tweaked alongside network settings for high-concurrency apps)
fs.file-max = 2097152

# ==============================================================================
# UNIVERSAL NETWORK TUNING (Performance & Stability)
# Optimized for: Low-Latency Gaming, VPNs, and Container Routing
# ==============================================================================

# --- 1. IP Forwarding (Routing Support) ---
# Enables IPv4 packet forwarding.
# Benefit: Necessary for functionality in Containers and VPN Mesh networks.
# Use Case: Required for Tailscale/ProtonVPN (Exit Nodes/Subnet Routers) and Docker/Podman containers.
net.ipv4.ip_forward = 1
Copy link
Member

@1Naim 1Naim Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Isn't this mainly for servers that are trying to act as routers? I don't see why this should be enabled for desktops.

Copy link
Contributor Author

@SoongVilda SoongVilda Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We already had that enabled; I just moved this to the network section to have everything in one place.

If you are not using Docker, KVM virtualization, or VPN, then that is useless, correct.

The host kernel must forward packets from the virtual interface (e.g., docker0, virbr0) to your physical interface (e.g., Ethernet or Wi-Fi).

Since Docker, KVM, or VPN is used on the desktop, I think we should keep it enabled as we have it.

Copy link
Member

@1Naim 1Naim Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't know about docker, but there's no need to enable this for KVM network access or using a VPN.

a. The virtual bridge for KVM works fine accessing the host network via dnsmasq without touching this setting.

b. No, you don't need this to use a VPN. It's needed if you're hosting a VPN server.

Copy link
Contributor Author

@SoongVilda SoongVilda Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Check this:

net.ipv4.ip_forward = 1, so I think this one should still be there. I also noticed we have some Tailscale users on Discord, and Tailscale requires this one if you want to act as exit node, but yeah, that's small amount users, mostly needed for Docker.

Copy link

@damachine damachine Dec 6, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That's very software-specific. Rather useless for the average desktop user. If the value is enabled, it applies not only to Docker but globally(for security reasons, this should remain disabled (set to 0) unless the software explicitly). With this enable (1) packages can then be forwarded. The computer acts as a router. I think if someone is using Docker or other specific (server) software, they'll already know what they need and will have to adjust it manually.

Hope that helps. KISS


# --- 2. MTU Probing (Connection Stability) ---
# Enables Path MTU Discovery.
# Benefit: Essential for connections involving encapsulation (VPNs) where the MTU is < 1500.
# Use Case: Fixes "black hole" connections and stalled transfers in WireGuard, OpenVPN, or mobile networks.
net.ipv4.tcp_mtu_probing = 1

# --- 3. TCP Timewait Reuse (Socket Efficiency) ---
# Allows reusing sockets in TIME_WAIT state for new connections.
# Benefit: Prevents "Port Exhaustion" during high connection turnover.
# Use Case: Improves performance for P2P apps (Syncthing/Torrents) and web servers.
net.ipv4.tcp_tw_reuse = 1
Copy link
Member

@1Naim 1Naim Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This also seems to be only beneficial for busy servers. IOW, doesn't have much effect if at all on desktops.

Copy link
Contributor Author

@SoongVilda SoongVilda Nov 21, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree this is a really edge-case parameter, which we could drop. It would be beneficial on desktop only for large torrenting.

Copy link

@damachine damachine Dec 6, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can`t say much about that but:
https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt

tcp_tw_reuse - INTEGER
	Enable reuse of TIME-WAIT sockets for new connections when it is
	safe from protocol viewpoint.
	0 - disable
	1 - global enable
	2 - enable for loopback traffic only
	It should not be changed without advice/request of technical
	experts.
	Default: 2

Hope that helps.


# --- 4. Input Packet Queue (Burst Handling) ---
# Increases the backlog queue for incoming packets.
# Benefit: Prevents losing packets when the interface receives data faster than the kernel can process it.
# Use Case: Smoother playback in streaming apps and better handling of high-speed downloads (Gigabit+).
net.core.netdev_max_backlog = 4096