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:
“...to improve the Antelope protocol (specifically) for node operators”.
Meetings take place every Wednesday from 14 UTC to 15 UTC (13 UTC to 14 UTC during daylight savings). For those wishing to learn the basics of EOS node operations, the EOS Network Foundation provides tutorials and documentation.
OVERVIEW
Discussion topics took on a different form in recent weeks. Not surprising given the impending launch of EOS EVM and the initial testing of Leap 4.0.0.
A few additions were made to the Special Nodes section of the ‘draft taxonomy of the roles that different antelope nodes play’ during the March 15 roundtable. The March 22 meeting was canceled. No additions to the draft document were made on March 29.
Key events impacting the community include:
EOS EVM undergoes a security audit in preparation for the April 14 launch
Leap 4.0.0 rc2 is live on the testnet
Community feedback took up much of the time for both the March 15 and 29 roundtables.
The ongoing Special Nodes discussion begins under the section entitled ‘Post Upgrade meetings’ on GitHub. Assurance was given that node taxonomy will continue. The ultimate goal is to improve the experience and potential of future node operations for both existing and new operators.
March 15: Leap 4.0.0, a Patch for 3.2.2, and Feedback
Watch the March 15 roundtable discussion on the ENF’s YouTube. Read the notes on EOS Nation (Leap) GitHub.
UPDATE
The meeting began with several updates. A patch (Leap 3.2.2) was released to fix SHiP stability. The Three-Strikes Rule garnered much feedback.
Expect updates for DUNE and CDT(contract developer toolkit) to follow soon after Leap 4.0.0 lands on the testnet.
Leap 4.0.0 will be discussed with the ENF in a future roundtable. What’s known thus far is that testing expects to begin in March. Be aware that there’s no short-term plan for a consensus upgrade until sometime in the Fall.
The technical discussion continued to populate the draft document. Detailed notes that relate to the ongoing Special Nodes and past discussions can be found there.
KEY TOPICS
Node-type items were postponed for next week. Highlights of the meeting include:
overview of the first Leap 4.0.0 RC (see AntelopeIO GitHub)
version lifecycles for node operators need
the future of Leap 3.2
What to Expect with the First Leap 4.0.0 Release Candidate and 3.2 Patch
Among the short-term focus for the Leap 4.0.0 RC is:
higher performance via multi-threading
reduced latency and faster block propagation
liter validation for block relays
Higher performance with multi-threading led to an interest in SHiP (State History Plugin). Within the node taxonomy draft document you’ll find:
“Console logging/print statements related to SHIP and DeepMind logging. Should there be limits?”
As well as, when to enable/disable SHiP for best practices. Parallelized execution also came up here.
Keep in mind that the node taxonomy draft document does not focus on Leap 4.0.0. At least, not yet. See AntelopeIO GitHub for bug fixes about SHiP stability.
Reducing latency and faster block propagation followed. Touched upon was scheduled-based auto peering for BPs to ensure receiving the last block with minimal delay. Auto-peering with scheduled proximal BP nodes (see AntelopeIO GitHub) seemed to stand out among the group.
Release notes about validation, a Prometheus Exporter Plugin, and more can be found on the AntelopeIO GitHub. Also, check for the latest Leap 4.0.0 RC and 3.2 patches. Ubuntu 18.04 vs. 20.04 and related security concerns were mentioned as the March 15 roundtable concluded.
Closing Feedback
The community was asked for feedback about how to focus support for the latest stable version of Leap moving forward.
The community should not expect a new (non-patch) release for 3.2. However, feedback is also wanted on the future of 3.2. One suggestion was to maintain (two) successive stable releases (i.e one for 4.0 and another for 3.2). Though, API changes still need to be considered.
OUTLOOK
Closing out the meeting was assurance that September/October remains the focus for a consensus upgrade.
March 22: CANCELED
March 22 was also the same week that the EVM code was completed and began live testing.
March 29: Feedback about Recent Updates
Watch the March 29 roundtable discussion on the ENF’s YouTube.
Given the nature of recent meetings, there’s more than the usual overlap of information. Readers may find the following document list convenient:
EOS Node Operator playlist
EOS Nation, Leap GitHub issues list
Be aware that no additional notes were added to the draft document on March 29.
UPDATES
As one might expect following the live testing and anticipated launch of EOS EVM, as well as testing Leap 4.0.0 rc2, updates quickly took over the meeting.
Expect a third release candidate within the next week or so. Spoiler Alert… shockingly rampant development got more intense with the first mention of a 5.0 consensus upgrade (for the coming Fall season). Also, expect a new CDT release in the weeks ahead and support for IPv4/IPv6.
CALL TO ACTION FOR BPS → BPs should sign the MSIG for creating the EVM account
KEY TOPICS
Enthusiasm led to lax discussions about:
blocks log
better Mac support
producer plugins
BP peering
About Blocks log, Mac, and Producer Plugins
The ‘blocks log’ discussion focused on participant clarity, particularly regarding operational efficiency. The AntelopeIO GitHub details ‘Blocks Log Splitting’. There’s a working plan to gauge user experience to better support Macs. Priorities for producer plugins include snapshots.
BP Peering
Subtopics related to BP peering feedback include:
improving connectivity
facilitating auto-configuration
configuration variability (e.g specifying BP name)
BP Proximity
effective outbound management
broadcasting and security
Release Notes and Closing Feedback
As anticipation heightens around a growing EOS ecosystem, the value of administrative and managerial efficiency took hold of the group. Among the concerns and interests were:
Alignment of information disseminated via dedicated documentation vs. release notes
How information and active upgrading can be designed to best serve the community and development performance
Suggestions include:
Providing examples and links within a documentation portal
Stable releases should align code with documentation
Release notes should not support a live, stable product
Conditions were identified where release notes alongside training may best serve in times of fast innovation. While node operators need to be aware of the latest features, the upgrading process can be improved to provide benefits that echo throughout the community.
Will there be a user-friendly dashboard for the Prometheus Exporter plugin? What about the mention of 5.0?
Prometheus remains a key focus. A consensus upgrade for Leap and mention of ‘5.0’ coincide with a September/October timetable.
Take the following list lightly. Not only does it forecast months ahead, but the discussion wasn’t prepared. Unexpected topics included:
Looking forward to 5.0 and optimized (compressed) block times
Releasing blocks earlier with percentages, offsets, and conservative values
Seamless, rather than 500 milliseconds, before starting new blocks
To keep up with the above comments, try starting with AntelopeIO GitHub Optimize block start time by huangminghuang.
Closing feedback debated how best to maintain old standards, clarity for an overload solution, exact time vs. a compression window, and dropping block offsetting.
OUTLOOK
The March 29 meeting ran a bit over. Seems as if everyone in the EOS community has their seat in the upright position. Not that anyone is awaiting in-flight, honey-roasted snacks. Rather, more eyes are becoming aware of the beautiful scenery just over the horizon.
EVM, a consensus upgrade, GameFi/metaverse funding, and more within the coming months make for an exciting time.
Sources & References