Skip to content

chore(dashmate): run Dash Core 24 by default - #5842

Merged
shumkov merged 14 commits into
v5.1-devfrom
shumkov/dashmate-core-24
Nov 12, 2026
Merged

shumkov merged 14 commits into
v5.1-devfrom
shumkov/dashmate-core-24

Conversation

@shumkov

@shumkov shumkov commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator

Issue being fixed or feature implemented

Dash Core 24.0.0 is released (dashpay/dashd:24 on Docker Hub) and the network runs it, but dashmate still sets up nodes with dashpay/dashd:23.

Middle of GitHub Stack #5847 (#5841 → this → #5843). On Core 24.0, the v5.1 DKG-safe stop rule refuses almost every masternode stop. #5841 fixes that, so this PR must not merge before it.

What was done?

  • The default Core image is dashpay/dashd:24.
  • A config migration keyed at 5.1.0-dev.1 moves configs still on dashpay/dashd:23 to the new default. It shares the key with the protxinfo whitelist migration from fix(dashmate): key the protxinfo whitelist migration to 5.1.0-dev.1 #5844, so both now live in one 5.1.0-dev.1 migration. dashpay/dashd:23 is the only Core 23 tag a release ever stored (the 3.0.1 migration and the default since), so any other image is the operator's choice and is left alone.
  • Only mainnet, testnet and the local preset (which new local networks are set up from) move. Core 24 discards the sporks Core 23 stored (the sporks.dat serialization version changed). On a local network or devnet whose nodes all move at once, no peer has them anymore, and dashmate signs local sporks only during setup. Existing local networks and devnets therefore keep Core 23, until they are set up again or their operator moves the image and re-signs the sporks.
    • The key is the first 5.1 release that will carry it. A 5.1.0 key would stamp prerelease testers' configs 5.1.0 and make later 5.1 prerelease migrations skip them.
  • docs/config/core.md named dashpay/dashd:22 as the default; it now names 24. The docs/update.md example, for the existing local network local_1, stays on dashpay/dashd:23.

Compatibility audit: dashmate on v5.1-dev was checked against Core v24.0.0 sources and release notes. Nothing it writes, calls or whitelists changes in a way that is not already handled:

  • dash.conf options and debug categories, and masternode start-up constraints;
  • dashd-rpc argument typing, protx through callWithParams, getdeploymentinfo, upcoming_dkgs, deprecatedrpc=service;
  • RPC whitelists, the Docker entrypoint, and the onion bind.

For operators (release notes):

  • Core 24 refuses to start with an invalid -debug category, or with -rescan / -minsporkkeys, in core.docker.commandArgs.
  • The first Core 24 start migrates insight/address/spent/timestamp indexes, which can take 20–40+ min. A restart --safe assumes the node is back within 6 blocks, so operators with these indexes should do this first start outside a DKG of their node.
  • Once Core 24 features activate, going back to dashpay/dashd:23 is not possible.

Note for later 5.1 migrations: semver orders beta < dev < rc. If 5.1 goes from dev.N to beta.1, as 4.2 did, a migration keyed 5.1.0-beta.N skips configs stamped by dev builds.

How Has This Been Tested?

Breaking Changes

None. Nodes on the stock Core 23 image move to Core 24 on the next dashmate update.

Checklist:

  • I have performed a self-review of my own code
  • My Rust, FFI, WASM, Swift or Kotlin changes follow the coding conventions
  • I have commented my code, particularly in hard-to-understand areas
  • I have added or updated relevant unit/integration/functional/e2e tests
  • I have added "!" to the title and described breaking changes in the corresponding section if my code contains any
  • I have made corresponding changes to the documentation if needed
  • If I added or changed GroveDB structure, I described it in the area's structure.rs, regenerated grovedb-structure.json, and checked the structure viewer link posted on this pull request
  • If I added or changed a system contract's document schemas, I checked the contract viewer link posted on this pull request

For repository code-owners and collaborators only

  • I have assigned this pull request to a milestone

🤖 Generated with Claude Code

@shumkov shumkov added this to the v5.1.0 milestone Nov 10, 2026
@coderabbitai

coderabbitai Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

Explain Change →

Caution

Review failed

The pull request is closed.

Pre-merge checks: Passed 1 · Inconclusive 1

Failed Failed checks: 1 inconclusive

Check name Status Explanation Resolution
Title check Pending This assessment has not completed. The check will be retried on the next pre-merge check run.
Passed Passed checks: 1 passed

Check name Status Explanation
Description Check Passed Check skipped - CodeRabbit’s high-level summary is enabled.
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

@thepastaclaw

thepastaclaw commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator

♻️ No change to this PR's diff since commit 36c9a02 (an empty commit, a merge of the base or a rebase onto it): that review stands — approved; @thepastaclaw review asks for a review of this commit (commit 2dc98da)

@thepastaclaw thepastaclaw left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Final validation — Phase 1 + Phase 2

The Core 24 default and exact-match migration are correct for public configurations, but the migration also repins existing local-network and devnet configurations. Core 24 does not retain the v23 runtime spork state across this transition, and dashmate only signs local sporks during setup, so updating and restarting an existing network can leave it without ChainLocks, InstantSend, or quorums. The update documentation also contains a stale Core 23 image example.

🔴 1 blocking | 💬 1 nitpick(s)

1 finding(s) not shown inline (the lines are not part of this PR's diff)

💬 Nitpick: Update example still reports Core 23 as the current image
packages/dashmate/docs/update.md:57

The JSON example for dashmate update --format=json still shows dashpay/dashd:23, while this change makes dashpay/dashd:24 the current shipped default. Keeping the old image in an example makes current update output appear inconsistent with the documented default.

"image": "dashpay/dashd:24"

source: glm-5.3-flash (phase1-reviewer: general)

Review provenance

Source: reviewer 1: glm-5.3-flash (agent: phase1-reviewer, role: general); reviewer 2: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 3: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); final verifier: gpt-6.1-sol (agent: sol-verifier, role: final-verifier)

  • Triage: low by gpt-6.1-sol (effort low) — This is a small, well-contained dashmate default-image and configuration-migration change with focused tests and no critical protocol, funds, cryptographic, or storage-consensus surface.
  • Phase 1 reviewers: glm-5.3-flash — general (completed, effort high); agent phase1-reviewer
  • Phase 1 model: glm-5.3-flash — zai quota: 5h 98% left, weekly 91% left; passed over gemini-3.8-flash-high (antigravity quota lookup failed)
  • Fresh verifier: gpt-6.1-sol — final-verifier; agent sol-verifier
  • Phase 2 reviewers: gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer
🤖 Prompt for all review comments with AI agents
These findings are from an automated code review (PastaClaw). How to work through them:
- Verify each finding against the code; fix only what is real, as narrowly as the finding.
- If a finding is wrong, decline it on its thread in one or two sentences; the bot answers once and stops.
- To defer one, link a tracking issue on its thread (or reply asking the bot to file one).
- Suggestions are optional: fix them forward in a later PR rather than pushing another round just for them.
- Address every current finding in one push: self-review the full diff first, and don't push piecemeal.
- Keep the PR's scope; don't grow it to satisfy hardening suggestions.
- Carried items were raised in an earlier round: read the thread first, and don't re-fix what you already fixed, declined or deferred there. Disputed and deferred items are already handled and are not listed.

In `packages/dashmate/configs/getConfigFileMigrationsFactory.js`:
- [BLOCKING] packages/dashmate/configs/getConfigFileMigrationsFactory.js:1876-1881: Migration repins existing local and devnet networks without restoring their sporks
  This migration applies to every configuration whose image is exactly `dashpay/dashd:23`, including existing local-network and devnet configurations. After the image changes to Core 24, the v23 spork state is not retained as active runtime state; however, `configureCoreTaskFactory` signs the local sporks only during `setup local`, not during `start`, `restart`, or `update`. An operator updating an existing local network or devnet therefore restarts the whole network without active ChainLocks, InstantSend, or quorum-related sporks, and there is no migration-time warning or re-sign step. Restrict this migration to public networks and the uninstantiated `local` preset, as the previous Core upgrade migration did, or implement an explicit spork re-sign flow and document it for existing local/devnet networks.

In `packages/dashmate/docs/update.md`:
- [NITPICK] packages/dashmate/docs/update.md:57: Update example still reports Core 23 as the current image
  The JSON example for `dashmate update --format=json` still shows `dashpay/dashd:23`, while this change makes `dashpay/dashd:24` the current shipped default. Keeping the old image in an example makes current update output appear inconsistent with the documented default.

@thepastaclaw thepastaclaw left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Re-review — Final validation — Phase 1 + Phase 2

The Core 24 default, scoped migration, and regression coverage are consistent, and no failed code checks are present on this head. One in-scope documentation example still reports Core 24 for an existing local configuration that the migration intentionally keeps on Core 23.

🟡 1 suggestion(s)

Approved: nothing here blocks the merge. The non-blocking findings are optional: fix them forward, in this PR or a follow-up, without waiting for another review round.

Review provenance

Source: reviewer 1: glm-5.3-flash (agent: phase1-reviewer, role: general); reviewer 2: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 3: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); reviewer 4: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 5: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); final verifier: gpt-6.1-sol (agent: sol-verifier, role: final-verifier)

  • Triage: low by gpt-6.1-sol (effort low) — This is a small, well-contained dashmate default and configuration-migration change with focused unit tests, and it does not modify a critical consensus, cryptographic, funds, or storage-migration surface.
  • Phase 1 reviewers: glm-5.3-flash — general (completed, effort high); agent phase1-reviewer
  • Phase 1 model: glm-5.3-flash — zai quota: 5h 88% left, weekly 88% left; passed over gemini-3.8-flash-high (antigravity quota lookup failed)
  • Fresh final gate: an independent Phase-2 review ran after iterative findings were reconciled
  • Fresh verifier: gpt-6.1-sol — final-verifier; agent sol-verifier
  • Phase 2 reviewers: gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer
🤖 Prompt for all review comments with AI agents
These findings are from an automated code review (PastaClaw). How to work through them:
- Verify each finding against the code; fix only what is real, as narrowly as the finding.
- If a finding is wrong, decline it on its thread in one or two sentences; the bot answers once and stops.
- To defer one, link a tracking issue on its thread (or reply asking the bot to file one).
- Suggestions are optional: fix them forward in a later PR rather than pushing another round just for them.
- Address every current finding in one push: self-review the full diff first, and don't push piecemeal.
- Keep the PR's scope; don't grow it to satisfy hardening suggestions.
- Carried items were raised in an earlier round: read the thread first, and don't re-fix what you already fixed, declined or deferred there. Disputed and deferred items are already handled and are not listed.

In `packages/dashmate/docs/update.md`:
- [SUGGESTION] packages/dashmate/docs/update.md:57: Do not show Core 24 for the existing local_1 update example
  This example runs `dashmate update --config local_1`, but the `5.1.0-dev.1` migration deliberately leaves existing local-network configurations such as `local_1` on `dashpay/dashd:23` because Core 24 discards the stored sporks and dashmate signs local sporks only during setup. The documented output therefore does not match the behavior for the named configuration. Either use a public configuration such as `testnet` in the command, or show `dashpay/dashd:23` for this existing local-network example.

Comment thread packages/dashmate/docs/update.md Resolved
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from f416721 to 4157832 Compare November 10, 2026 17:13

@thepastaclaw thepastaclaw left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Re-review — Final validation — Phase 1 + Phase 2

The current head only changes the existing local_1 update example to show Core 23, matching the migration's deliberate preservation of already-created local networks. No new in-scope issues were identified; the available CI snapshot contains no failed build, test, or lint checks.

🔴 0 blocking | 🟡 0 suggestion(s) | 💬 0 nitpick(s)

Review provenance

Source: reviewer 1: glm-5.3-flash (agent: phase1-reviewer, role: general); reviewer 2: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 3: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); final verifier: gpt-6.1-sol (agent: sol-verifier, role: final-verifier)

  • Triage: low by gpt-6.1-sol (effort low) — The diff is a small, contained default-image update and configuration migration with explicit network and image filters and focused tests, not an intricate change to a critical surface.
  • Phase 1 reviewers: glm-5.3-flash — general (completed, effort high); agent phase1-reviewer
  • Phase 1 model: glm-5.3-flash — zai quota: 5h 87% left, weekly 87% left; passed over gemini-3.8-flash-high (antigravity quota lookup failed)
  • Fresh verifier: gpt-6.1-sol — final-verifier; agent sol-verifier
  • Phase 2 reviewers: gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer
🤖 Prompt for all review comments with AI agents
These findings are from an automated code review. Verify the current code and confirm that no unresolved issues remain.

No unresolved findings remain from the prior review on this head.

@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch 2 times, most recently from 9a9952f to 3e70b3e Compare November 10, 2026 20:46
@shumkov
shumkov requested a review from ktechmidas November 11, 2026 10:44
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from 3e70b3e to 710b9ed Compare November 11, 2026 10:52
shumkov and others added 7 commits November 11, 2026 17:53
`dashmate stop`, `restart` and `--safe` decide when stopping a masternode
cannot cost it a PoSe penalty for missing DKG messages. On Core 24.0 the
rule never let most masternodes stop: it treats any non-zero
`quorum dkginfo` `active_dkgs` as a running session once Core reports
`upcoming_dkgs`, but Core 24.0 still counts the DKG worker's session
records, which linger until the next round of the same quorum type and
index. Only Core 24.1 derives the count from membership. With Core 24.0
plain `stop` refused about 94% of the time on mainnet and `--safe`
waited hours to days.

One rule now serves Core 23, 24.0 and 24.1:

- A DKG this node is a member of (or whose membership Core cannot
  determine) blocks from 6 blocks before its start. Core 24 reports
  membership in `upcoming_dkgs`; without it, any DKG within 6 blocks
  blocks, as before.
- `quorum dkgstatus` records block from a session's start to the end of
  its Commit phase (5 * dkgPhaseBlocks), measured from each rotated
  quorum index's own start. Rotated `llmq_60_75` sessions blocked for 42
  blocks from their index's start (the cycle's shared mining window,
  counted from the wrong base); they now block for 10, as Core 24.1
  counts them.
- A record that has reached its next round but that Core has not yet
  replaced blocks as that new round.
- Records are read on every Core version: on 24.1 a mid-session
  ProUpServTx clears the proTxHash, so `active_dkgs` drops to 0 while
  the session runs.
- Core 24.1 and later: a non-zero `active_dkgs` blocks.

Core older than 24.1 records a session only after the start block
connects, so a "safe" result taken right at a DKG start could be wrong.
Each check now reads the chain tip before and after the DKG RPCs and is
discarded if it moved. On Core older than 24.1 a safe result is
confirmed 15 s after the first check finished, provided the chain moved
by at most 6 blocks and did not replace the first check's tip. `--safe`
also remembers member DKGs it saw announced, which covers the gap where
Core has erased last round's record but not yet created the new one.
Core in initial block download takes no part in DKGs and may stop.
Plain stop retries an unsettled check up to 3 times, and reports
unreadable Core data with its own error instead of suggesting --safe.

`group restart --safe` declared its flag as a plain object, which oclif
parsed as false, so the restart never waited.

Cost: on Core older than 24.1 every plain stop or restart of a masternode
that is safe to stop takes about 15 s longer.

Tests would have caught this in CI. Run against the previous code, these
fail: ✖ before the fix, ✔ after:
- should stop waiting on Core 24.0 when active_dkgs counts only finished sessions
- should end a rotated DKG session once its own Commit phase is over
- should keep waiting at a DKG start block while Core still shows the previous round
- should not stop at a DKG start that Core has not recorded yet
- should not trust DKG reads that saw a newer tip than the height read
- should keep waiting on Core 24.1 while a session runs after the proTxHash was cleared
- should parse --safe so the restart waits for a safe DKG window

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On Core older than 24.1, a check that begins at a DKG's start block can
fall into the gap where Core has erased the previous round's record but
not yet recorded the new session, and that gap has no upper bound.
`quorum dkgstatus` lists every enabled quorum with the start of its
current session, so a session starting at the tip without a record now
blocks that block. `--safe` remembers DKGs Core announced without this
node, and a confirming check reuses the first check's announcements, so
their start blocks do not wait for a record that never comes. A
`dkgstatus` answer without quorum connections is unreadable.

The confirming check now reads the active chain's block at the first
check's height inside its consistent snapshot. A reorg that replaced the
first tip and ended up to 6 blocks higher passed the old same-height
hash comparison.

Group commands stop a local network's miner before its masternodes.
With `--safe`, now effective for `group restart`, the masternodes then
waited for blocks that could never come. Once a local network's miner is
stopped, later masternodes in the run skip the DKG check: with every
masternode going down together, their DKG produces no commitment to
penalize them in.

Core in initial block download stays safe to stop when blocks connect
between the reads, and a Core version that is not an integer is now an
error instead of being taken for an old Core.

Tests would have caught these: ✖ before the fix, ✔ after:
- should not stop at a DKG start reached in Core's record gap, however long it lasts
- should refuse a plain stop at a DKG start that Core has not recorded yet
- should report unsettled when a reorg replaced the first check's tip and the chain grew
- should not wait for DKGs once the local network's miner has stopped
- should treat Core in initial block download as safe even while blocks keep connecting
- should refuse to guess when Core reports no integer version
- should pass on what it knows about DKGs this node is not part of

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…le window

On Core older than 24.1, a session's start only blocked its start block.
Core records a session after computing its members, which has no time
bound, so a check at the next block without a record could not tell a
non-member from a member Core had not recorded yet. A current session
without a record, from `quorum dkgstatus` quorum connections, now blocks
until its Commit phase ends, unless Core announced it without this node.
A plain stop is refused for the first blocks of such a session; `--safe`
does not wait for sessions it saw announced without this node.

Remembered non-memberships held across a reorg. A replacing branch has
another work block and may make this node a member, so each poll checks,
inside its consistent snapshot, that the previous poll's tip is still on
the chain, and a check that finds it replaced, or a confirmation that
finds the first check's tip replaced, reports a reorg and drops them.

The stopped-miner exemption applied to every config in the run. A group
may hold configs of several networks, so it now covers only local-network
configs in the stopped miner's group.

Tests would have caught these: ✖ before the fix, ✔ after:
- should keep blocking after the start block while Core still has no record
- should not stop in a session Core has not recorded when waiting starts after its start block
- should forget DKGs announced without this node when a reorg replaces their chain
- should report a replaced chain when the previous check's tip is gone
- should still check a masternode of another network after a local miner stopped

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…session ends

Now that a session Core has not recorded blocks for its whole window, the
wait must keep a DKG it saw announced without this node until that
window ends, not only until its start block, or a non-member waits the
session out anyway.

Test would have caught this: ✖ before, ✔ after ("should not hold a
session it saw announced without this node after its start block").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`--safe` stopped a node as soon as Core reported initial block download,
even when it had seen a DKG this node is a member of announced. A Core
restarted into initial block download may finish syncing in time for
that DKG, so a remembered DKG now blocks while Core syncs too.

An announced DKG of a quorum type dashmate does not know, that may
include this node, is now unreadable rather than a blocker with no end:
without its Commit phase nothing says when it stops blocking, and a
remembered one would hold a `--safe` wait forever. Unknown types in
session records and quorum connections were already unreadable.

Tests would have caught these: ✖ before, ✔ after:
- should keep blocking a DKG it saw announced when Core restarts into initial block download
- should not read a DKG of a type dashmate does not know that may include this node

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Core in initial block download was safe to stop without looking at its
DKG state. It runs no DKG session until synced, so session records do not
apply, but a Core only a little behind can finish syncing in time for the
DKGs it announces; announced, imminent and unreadable DKGs now still
block while it syncs. When Core restarts into initial block download
during a confirmation, the first check's DKGs are kept.

The `--safe` wait read Core's version once. Core replaced between polls
without a failed call kept the old version's rule, so a Core 24.0 that
counts records was read as Core 24.1 that counts members. The version is
now read on every poll.

Tests would have caught these: ✖ before, ✔ after:
- should still block a DKG this node is in while Core syncs
- should keep the first check's DKGs when Core restarts into initial block download during confirmation
- should judge each poll by the Core version answering it

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An announced DKG's start was rebuilt as the tip plus Core's
`blocksUntilStart`, and `next_dkg` was counted from the tip read after
`dkginfo`. While Core syncs, a snapshot is kept although the tip moves,
so a block connecting after `dkginfo` put the start one block late and
could let a stop through right at the 6-block margin. Starts now come
from Core's absolute `quorumHeight`, as do the keys of DKGs announced
without this node, and `next_dkg` counts from the lowest tip `dkginfo`
may have seen.

`--safe` remembered only announced DKGs. A Core restarted between polls
keeps no record of a running session, and Core 24.1 without its
proTxHash yet reports no count for it either, so a session the wait had
already seen could be forgotten mid-way. Every blocker with a known range
is now remembered until it ends.

Tests would have caught these: ✖ before, ✔ after:
- should place an announced DKG at the start Core gives
- should count next_dkg from the earliest tip dkginfo may have seen
- should keep blocking a running session after Core restarts and forgets it
- should block when an entry is missing quorumHeight

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from 710b9ed to c1c094f Compare November 11, 2026 10:58

@thepastaclaw thepastaclaw left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Re-review — Final validation — Phase 1 + Phase 2

At c1c094f, the local_1 documentation example correctly retains Core 23, and no new actionable defects were confirmed. The proposed mixed-image finding is prevented by the setup command's existing-group guard. Static review only: current CI shows JS build, lint, local-network E2E, and testnet-Evonode E2E passing; Dashmate unit tests and several downstream suites remain pending, with no failed code checks reported.

🔴 0 blocking | 🟡 0 suggestion(s) | 💬 0 nitpick(s)

Review provenance

Source: reviewer 1: glm-5.3-flash (agent: phase1-reviewer, role: general); reviewer 2: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 3: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); final verifier: gpt-6.1-sol (agent: sol-verifier, role: final-verifier)

  • Triage: low by gpt-6.1-sol (effort low) — The diff is a small, contained default-image update and configuration migration with explicit network and image filters and targeted tests, not a large or intricate change to a critical surface.
  • Phase 1 reviewers: glm-5.3-flash — general (completed, effort high); agent phase1-reviewer
  • Phase 1 model: glm-5.3-flash — zai quota: 5h 99% left, weekly 81% left; passed over gemini-3.8-flash-high (antigravity quota lookup failed)
  • Fresh verifier: gpt-6.1-sol — final-verifier; agent sol-verifier
  • Phase 2 reviewers: gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer
🤖 Prompt for all review comments with AI agents
These findings are from an automated code review. Verify the current code and confirm that no unresolved issues remain.

No unresolved findings remain from the prior review on this head.

…ewer tip

While Core syncs, a snapshot is kept although the tip moves during the
reads. Core lists a DKG in `quorum dkginfo` once its work block, 8 blocks
before its start, is mined, so a list read at a lower tip lacks DKGs that
the 6-block restart margin of the later tip may need. A snapshot whose tip
moved by more than the 2 blocks between the two is now unsettled.

Test would have caught this: ✖ before, ✔ after ("should not read Core's
announcements as covering a tip that moved too far past them").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from c1c094f to 36c9a02 Compare November 11, 2026 15:43
@shumkov

shumkov commented Nov 11, 2026

Copy link
Copy Markdown
Collaborator Author

/self-reviewed

@thepastaclaw thepastaclaw left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Re-review — Final validation — Phase 1 + Phase 2

At 36c9a02, the Core 24 default and selective configuration migration are consistent with the regression tests and existing-local update example. One documentation suggestion remains: publish the Core 24 upgrade precautions and private-network upgrade procedure in the operator-facing update guide. Static validation only; Dashmate CI tests are skipped, and the failed “Build JS packages / Build JS” check ends with Ubuntu mirror DNS failures and apt exit code 100, unrelated to this change.

🔴 0 blocking | 🟡 0 suggestion(s) | 💬 0 nitpick(s)

Approved: nothing here blocks the merge. The non-blocking findings are optional: fix them forward, in this PR or a follow-up, without waiting for another review round.

Follow-up candidates (optional) (1)

New non-blocking observations on code unchanged since the last full review. No inline threads are opened for them and later reviews do not repeat them: fix forward or track separately if they are worth it.

🟡 Suggestion: Publish the Core 24 operator guidance in the update guide
packages/dashmate/docs/update.md:86

The README links operators to this update guide, but it contains no Core 24 upgrade section. The migration now distinguishes public networks from existing local networks/devnets, while the PR description documents additional startup precautions, including the first-start index migration and its interaction with restart --safe. These details matter operationally: the safe-stop calculation reserves only six blocks before a DKG, and local sporks are signed during setup, not ordinary restart or update. Add a Core 24 section explaining which configurations move automatically, which images are preserved, the first-start and removed-argument precautions, and how to upgrade an existing private network by moving its nodes together and re-signing the supported sporks, or recreating the network. The corresponding update-guide section in PR #5827 provides reusable private-network instructions; include the additional first-start precautions documented in this PR as well.

source: glm-5.3-flash (phase1-reviewer: general)

Review provenance

Source: reviewer 1: glm-5.3-flash (agent: phase1-reviewer, role: general); reviewer 2: gpt-6.1-sol (agent: phase2-reviewer, role: general); reviewer 3: gpt-6.1-sol (agent: phase2-reviewer, role: architecture-layering); final verifier: gpt-6.1-sol (agent: sol-verifier, role: final-verifier)

  • Triage: low by gpt-6.1-sol (effort low) — The diff is a small, contained default-image update and config migration with straightforward image and network guards plus focused tests, not an intricate change to a critical surface.
  • Phase 1 reviewers: glm-5.3-flash — general (completed, effort high); agent phase1-reviewer
  • Phase 1 model: glm-5.3-flash — zai quota: 5h 78% left, weekly 76% left; passed over gemini-3.8-flash-high (antigravity quota lookup failed)
  • Code unchanged since the last final review (c1c094f6): 1 new suggestion(s) listed as follow-up candidates, not posted inline
  • Fresh verifier: gpt-6.1-sol — final-verifier; agent sol-verifier
  • Phase 2 reviewers: gpt-6.1-sol — general (completed, effort medium); agent phase2-reviewer, gpt-6.1-sol — architecture-layering (completed, effort medium); agent phase2-reviewer
🤖 Prompt for all review comments with AI agents
These findings are from an automated code review. Verify the current code and confirm that no unresolved issues remain.

No unresolved findings remain from the prior review on this head.

…s chain

While Core syncs, its tip can move several blocks during one read, which
made every plain-stop attempt unsettled, so a syncing masternode could not
be stopped without --force. Core in initial block download takes no part
in DKGs and stays in it only while connecting blocks over 6 hours old, so
how far its tip moved during the reads no longer discards them.

What does matter is whether the chain read last still has the work blocks
Core worked out membership from: a branch switched during the reads may
put this node in a DKG Core said it was not in. A read whose tip changed
is now unsettled unless every work block `quorum dkginfo` names is on that
chain. The previous check's tip is also looked up after the DKG reads, on
the chain the check reports.

Test would have caught this: ✖ before, ✔ after
("should not refuse a Core that connects several blocks during the reads",
"should report unsettled when the block Core worked out membership from is
replaced", "should report a replaced chain when the previous check's tip
is replaced during the reads").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from 36c9a02 to 0d5623d Compare November 11, 2026 18:04
When the read confirming a safe result was discarded, as when a block
connects during it, the check came back with no height and no blockers.
`--safe` remembers DKGs only from a check with a height, so it forgot
the DKG of this node the first read announced, and could stop the node
once Core no longer showed it. Such a check now reports the first read,
unsettled, so its announced DKGs are remembered.

Test would have caught this: ✖ before, ✔ after ("should keep what a
check announced when its confirming read is discarded").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from 0d5623d to 6c9e4dd Compare November 11, 2026 20:03
shumkov and others added 4 commits November 12, 2026 03:59
…e confirming

When Core older than 24.1 changed its session records during the read
confirming a safe result, the check came back unsettled with no
blockers. `--safe` then forgot the DKG of this node the first read
announced, and could stop the node once Core no longer showed it. Such
a check now carries the first read's blockers.

Test would have caught this: ✖ before, ✔ after ("should keep what a
check announced when Core changes its session records during the
confirming read").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New nodes get `dashpay/dashd:24`. A config migration keyed at 5.1.0-dev.1
moves every config still on `dashpay/dashd:23`, the only Core 23 tag a
release stored, to the new default. Any other image is the operator's
choice and is left alone.

Dashmate already handles Core 24's RPC and config changes: local setup,
status, the Drive consensus whitelist (`getdeploymentinfo`), the onion
bind and the DKG-safe stop rule. Nothing else dashmate writes, calls or
whitelists changes in Core 24.0.0.

The config docs listed `dashpay/dashd:22` as the default; they now name
24.

Test would have caught a missing migration: ✖ before, ✔ after
("should move a node on the stock Core 23 image to Core 24 and keep an
operator image").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Core 24 discards the sporks Core 23 stored. On a local network or devnet
whose nodes all move to Core 24 at once no peer has them anymore, and
dashmate signs a local network's sporks only while setting it up, so the
Core 24 migration would leave such a network without ChainLocks,
InstantSend or quorums after the next restart. It now moves only public
networks and the `local` preset new local networks are set up from;
existing local networks and devnets keep Core 23 until their operator
sets them up again or moves the image and re-signs the sporks.

The update docs example showed the Core 23 image.

Test would have caught this: ✖ before, ✔ after ("should keep a local
network or devnet already set up on Core 23").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The example updates `local_1`, an existing local network, which the Core
24 migration leaves on `dashpay/dashd:23`.

Docs only; no test.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shumkov
shumkov force-pushed the shumkov/dashmate-core-24 branch from 6c9e4dd to 2dc98da Compare November 11, 2026 21:01
@shumkov

shumkov commented Nov 12, 2026

Copy link
Copy Markdown
Collaborator Author

/self-reviewed

@shumkov
shumkov removed this pull request from stack #5847 November 12, 2026 01:42
Base automatically changed from shumkov/dashmate-dkg-safe-stop to v5.1-dev November 12, 2026 01:42
@shumkov
shumkov merged commit 1f7804c into v5.1-dev Nov 12, 2026
32 checks passed
@shumkov
shumkov deleted the shumkov/dashmate-core-24 branch November 12, 2026 01:42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants