How Everstake Optimizes Their Agave Nodes on Cherry Servers

Agave is a Solana-based blockchain with similar infrastructure demands—high throughput, fast block times, and performance-critical consensus. Running an Agave validator or RPC node isn’t just about launching a server. It’s about deep-level system tuning and hardware control.

Whether you're supporting the network as a validator or powering dApps via RPC, you’ll need fast NVMe storage, optimized kernel parameters, and reliable CPU performance. That’s where Cherry Servers comes in.

We’ve partnered with Everstake, the leading staking provider, to test, tune, and deploy Agave nodes across our bare metal infrastructure. This guide includes real-life tips and recommendations from Everstake’s DevOps team on how to optimise Agave nodes on Cherry Servers.

Check Everstake’s initial Solana node guide for detailed setup instructions.

#About Everstake

Everstake is a non-custodial staking and validator platform founded in 2018, trusted by over 800,000 users across more than 85 blockchain networks. With enterprise-grade infrastructure and ~99.9% uptime, Everstake empowers both individual and institutional clients to earn rewards securely while supporting decentralization.

The team consists of experienced engineers and DevOps professionals who contribute to ecosystem development, governance participation, and tooling for leading protocols. Everstake is also a key contributor to public blockchain education and ecosystem grants.

#How to Optimize Your Agave Node on Cherry Servers

Cherry Servers is a European-based cloud infrastructure provider offering bare-metal cloud solutions. Unlike standard virtualized setups, Cherry Servers allows complete control over dedicated hardware, making it an ideal choice for blockchain validators and RPC node operators.

Everstake chose Cherry Servers to support their validator operations, seeking a provider that delivers reliable hardware performance, rapid NVMe storage, and the flexibility to optimize systems at a low level, all without the complexities of virtualization.

If you've already gone through the Agave official setup docs, the below steps detail how to make your infrastructure faster and more stable. We’ll walk through practical system- and hardware-level optimizations for running Agave nodes on Cherry Servers.

#Step 1. Choose the Right OS & Kernel

Ubuntu 22.04 is widely used, but kernel 6.5+ is recommended for features like amd_pstate . If you’re using 22.04, enable the HWE stack to access newer kernels. Just beware of network interface changes post-upgrade—plan for them with netplan or kernel flags.

Upgrading to kernel 6.5+ can change network interface names, potentially causing connectivity loss after reboot. Add net.ifnames=0 biosdevname=0 to GRUB config or adjust Netplan using MAC addresses.

#Step 2. Tune BIOS/UEFI Settings

Make sure CPU boost is enabled, C-States are disabled (for latency-sensitive workloads), and fan profiles are set for high cooling. Cherry’s servers are typically shipped ready, but validating your motherboard config is still best practice.

#Step 3. Optimize Disk Layout

Use 3 separate Gen4 NVMe SSDs for:

• OS
• Ledger
• Accounts + Snapshots

This avoids I/O contention, especially during snapshot creation. Some operators also use RAID for accounts to prevent single-disk bottlenecks.

Important: never place Ledger and Accounts on the same drive, this can cause performance degradation under load. Use ext4 with mount options like noatime,nodiratime,data=writeback,barrier=0 for improved performance (with acceptable risk).

#Step 4. Apply Kernel and sysctl Tweaks

Use tuned sysctl parameters for TCP, UDP, memory, and file descriptor limits. These tweaks reduce network latency, increase buffer sizes, and stabilize memory behavior under heavy transaction load. Always test changes on staging nodes first.

#Step 5. Set the Right CPU Governor

Use the performance governor to lock cores at max frequency. For AMD CPUs on kernel 6.5+, enable amd_pstate=active for better scaling. Changes can be made persistent via GRUB parameters.

If supported, consider using amd_pstate in passive or guided modes for more advanced scaling. Enable CPPC in BIOS if needed.

Bonus: Compile from Source for Full Control Agave is moving away from distributing binaries, so compiling from source with lto=fat and target-cpu=native is the future. This approach gives you total control and sometimes faster performance, especially on modern CPUs.

Want the full deep-dive? Check out the extended guide on Everstake’s blog.

#Why Everstake Optimizes Their Agave Nodes on Cherry Servers

Everstake’s DevOps team chose Cherry Servers to run their Agave validators thanks to BIOS-level access, consistent NVMe throughput, and the ability to apply custom system tweaks. Their production nodes now run on our infrastructure with reduced downtime, stable sync, and maximum control.

“We were able to apply hardware and kernel-level tuning directly, which wouldn’t be possible on many cloud platforms. That made a real difference in node stability.” – Everstake DevOps

About Cherry Servers

Cherry Servers is a Web3-friendly European-based premium bare metal cloud provider. Ready to take your Agave infrastructure to the next level? Reach out and we’ll help you tailor the ideal setup based on real-world performance. Our team is here to help you get started with the right setup!