Author: Marco González

Editor: Randall Roland

Node operators, Antelope core developers, and community members get together each week to discuss the captivating questions of the day. The primary objective of each Node Operator Roundtable is:

Meetings occur every Wednesday from 14 UTC to 15 UTC (13 UTC to 14 UTC during daylight savings). The EOS Network Foundation provides tutorials and documentation for those wishing to learn the basics of operating an EOS node (and more).

Below is a list of the three roundtables briefed in this bi-monthly summary:

Look for additional meeting notes and comments on GitHub. Video recordings can be found on the ENF’s YT.

Leading web3 development means offering the best peer-to-peer experience. P2P improvements were the topic of the May 17 Node Operator Roundtable.

As usual, the roundtable started with a few minutes of technical updates.

Prior to the P2P discussion, quick feedback about EVM was taken. The P2P discussion centered around efficient peering, syncing, and snapshots.

EOS EVM feedback was quickly heard. Comments would be looked into and answered at a later meeting or via community socials. Among the comments made were:

Two topics stood out among the P2P improvements discussion:

Participants hope to see improvement in peering and related features. Focusing on fundamental improvements with approachable goals moved the conversation forward. Robust auto-discovery for known peers emerged as a potential path forward.

Underlying problems include:

Examination of the WAX blockchain may help solve pressing issues.

By far, the most comments fielded during the May 17 roundtable were about snapshots. What follows is a brief overview of comments and issues. Watch the recording for details.

The snapshot topic began by asking who most recently used the feature. Most participants said they use a snapshot about once per day. Reasons given included RAM and rebooting (for those running big nodes). Snapshots generally ran predictably for ‘good’ setups.

WAX node operators tend to run relatively large setups. The need to coordinate with WAX blockchain maintenance happens weekly. The process generally goes smoothly.

Restoring via snapshot generally takes 15 minutes, but several factors can extend that time to several hours. This topic brought on a lively conversation. For those seeking details, the snapshot discussion takes place during the last third of the meeting.

The May 24 roundtable began with guiding questions that were on-topic and of general interest.

The only update for today is:

Guiding topics posed to participants included:

The main discussion began with the host inviting node operators to share their experiences. The first topics posed sought to understand normal node operator maintenance, the processes involved, and how many ‘build from source’. The next question involved automation for nodeos. The meeting concluded with a follow-up of the May 17 roundtable discussing P2P improvements.

Node operators shared their insights into ‘build from source’ vs. a package manager. A package manager (e.g. activate install) might best suit less dedicated operators. Efficiency seemed universally welcomed.

Complex operations can involve many factors. Maintaining options seems prudent. Custom builds and configurations were a couple of examples. The discussion went as far as binaries, other chains, and custom app packages.

Following questions about automation, the challenges involved in node deployment took over the discussion.

Spinning up a useful node can be more complicated than it might first seem. Specifically mentioned was the gap between a ‘vanilla’ Docker image and deployment. For those less focused on customization, an app package can complement the use of Docker.

Cleos compiling was brought up for operations running with more options. Topics include were:

The explanation given for the need for Apple was that it’s challenging to run a virtual machine on nodeos. Also, many commands no longer work.

Closing out the automation discussion were binaries and nodeos.

As mentioned in the May 17 P2P discussion, snapshots brought on lively discourse. The May 24 meeting closed out with a more focused P2P follow-up discussion.

The snapshot feature is generally meant for disaster recovery type operations rather than its current high-frequency use. Node operators went on to describe their operations and encounters.

Optimizing/redesigning nodeos, temporary file storage, trade-offs (i.e. performance vs. database), traffic, echoing, and the need for better communication between the node operator community and Antelope developers were all discussed. The snapshot topic again drew enthusiastic responses.

If there was a takeaway it might be that identifying between tech and behavioral (of peer networks) issues might well serve the community and future development.

The final roundtable of the month was an open discussion primarily about RPC node infrastructure for EOS EVM.

No updates were mentioned for May 31.

Interesting topics on Telegram for the week include:

With no pressing topic, the floor was open to the community. EOS EVM infrastructure was brought up and turned into an extended discussion.

After several minutes, the host summarized the topic in the following way,

Both current users and developers of EVM expect centralized RPC nodes. It’s a dynamic that’s proven highly performant and streamlines development.

A key reason is that Ethereum developed alongside dapps running on compatible chains. Unlike layer-2 solutions, EOS EVM built a road TO Ethereum. Thus, where layer-2 chains model Ethereum technology, EOS offers more function, stability, security, and performant opportunities.

Ethereum is both older and more evolved with more financial capital than EOS. Extensive libraries and development options make these facts apparent. As EOS works to innovate its unique advantages into the ecosystem, RPC node infrastructure plans to remain centralized.

Currently, the ENF runs the RPC infrastructure. Expect to hear about an ENF partnership (e.g. Infuria and Alchemy) that eventually transitions the RPC node infrastructure from ENF management.

Running one’s own EVM RPC node outside of the current and planned centralized framework will be possible. The ENF will also provide information about how to run an independent EVM node. However, decentralized operations won’t have access to networking advantages such as marketing and infrastructure.

Some topics briefly mentioned as the meeting closed include:

Everyone is encouraged to upgrade from 3.2 to 4.0 and share their experiences.

Roundtables continue to tackle issues that make for more efficient and refined node operations as Antelope and EVM development continues toward Leap 5.0. Node operators may play more vital roles in onboarding community members. Communication, productive feedback, overall performance, and support for innovation all show uptrends with each passing meeting.