BBR is a TCP congestion-control algorithm that estimates bandwidth and round-trip time. It can improve throughput on long-distance or high-latency TCP connections, but it is not a universal speed boost. If the bottleneck is CPU, storage, a bandwidth cap, a poor route, or UDP traffic, BBR may make little difference.
Check support first
uname -r
sysctl net.ipv4.tcp_available_congestion_control
sysctl net.ipv4.tcp_congestion_control
sysctl net.core.default_qdisc
Linux 4.9 or newer is generally required. Continue only if the available algorithms include bbr. If BBR is missing, editing sysctl configuration will not help; upgrade to a supported kernel first.
Enable BBR safely
Back up the configuration and check for conflicting entries:
sudo cp /etc/sysctl.conf /etc/sysctl.conf.bak
grep -nE 'net.core.default_qdisc|net.ipv4.tcp_congestion_control' /etc/sysctl.conf
If the keys are absent, add:
net.core.default_qdisc=fq
net.ipv4.tcp_congestion_control=bbr
Then load the configuration:
sudo sysctl -p
Modern Ubuntu, Debian, CentOS Stream, Rocky Linux, and AlmaLinux installations usually need no extra steps when the running kernel supports BBR. CentOS 7 often has an older default kernel, so check the actual kernel.
Verify the active setting
sysctl net.ipv4.tcp_congestion_control
sysctl net.core.default_qdisc
lsmod | grep bbr
The important result is net.ipv4.tcp_congestion_control = bbr. Some systems may not show an obvious lsmod entry even when BBR is active.
Measure instead of guessing
Compare before and after results using the same destination, file size, time window, and route. Look at sustained throughput, rsync or SCP stability, retransmissions, and performance over high-latency links. A single page refresh is not a useful benchmark.
If sysctl -p reports an invalid argument, recheck the available algorithms and look for duplicate settings in /etc/sysctl.conf and /etc/sysctl.d/. If the value changes after reboot, another sysctl file, cloud-init, a control panel, or an optimization script may be overriding it.
Roll back
sudo cp /etc/sysctl.conf.bak /etc/sysctl.conf
sudo sysctl -p
The practical rule: verify kernel support, back up the configuration, enable BBR, confirm the active value, and measure under repeatable conditions.
The original GSVPS guide includes distro-specific commands and a status-check script.
Top comments (1)
🧠 What BBR is (quickly)
TCP BBR is a congestion control algorithm designed to improve throughput and reduce latency by estimating bandwidth + RTT instead of reacting only to packet loss.
It works best on modern Linux kernels (typically 4.9+, ideally newer).
⚙️ 1. Check if BBR is available
Run:
sysctl net.ipv4.tcp_available_congestion_control
If you see:
reno cubic bbr
👉 BBR is already supported.
📦 2. Load BBR module (if needed)
Check module:
lsmod | grep bbr
If empty, load it:
sudo modprobe tcp_bbr
To persist after reboot:
echo "tcp_bbr" | sudo tee /etc/modules-load.d/bbr.conf
🚀 3. Enable BBR (system-wide)
Edit sysctl config:
sudo nano /etc/sysctl.conf
Add:
net.core.default_qdisc=fq
net.ipv4.tcp_congestion_control=bbr
Apply immediately:
sudo sysctl -p
🔍 4. Verify BBR is active
Run:
sysctl net.ipv4.tcp_congestion_control
Expected output:
net.ipv4.tcp_congestion_control = bbr
Also verify queue discipline:
sysctl net.core.default_qdisc
Expected:
fq
🧪 5. Optional live confirmation
ss -tin
Look for bbr in congestion control details.
🧩 Key takeaway
BBR must be supported by kernel
Must be enabled via sysctl
Works best with fq qdisc
Verification is always done via sysctl