fix(SEV-046): bump deploy SOL floor 5→20 (empirical 12.5 SOL needed) by alrimarleskovar · Pull Request #393 · alrimarleskovar/RoundFinancial

alrimarleskovar · 2026-05-18T13:55:51Z

TL;DR

Rehearsal-1f finally got past every workflow bug (PRs #389, #390, #391, #392) and surfaced the real cost: anchor deploy of 4 programs needs ~12.5 SOL, not the ~3-4 SOL my original docs claimed. Bumps the workflow floor 5 → 20 SOL.

What rehearsal-1f surfaced

$ anchor deploy --provider.cluster devnet
Deploying program "roundfi_core"...
Program path: /home/runner/.../target/deploy/roundfi_core.so...
Error: Account 5ZpFtJePb2hGKhG9RJ6Fdwmo5y8wuwKXZZcKttoN1Jgo has
       insufficient funds for spend (6.13730712 SOL) + fee (0.00438 SOL)

Two important facts here:

anchor deploy actually invoked — keypair found, build successful, balance check passed. All previous PRs' fixes work.
The deploy needs more SOL than the workflow's 5 SOL floor lets through (the check passed at exactly 5 SOL, then ran out mid-deploy of the first program).

Empirical per-program cost

Program	Bytecode	2× rent-exempt
`roundfi_core`	~880KB	~6.14 SOL
`roundfi_reputation`	~370KB	~2.6 SOL
`roundfi_yield_kamino`	~294KB	~2.0 SOL
`roundfi_yield_mock`	~254KB	~1.8 SOL
TOTAL	~1.8MB	~12.5 SOL + tx fees

Fix

.github/workflows/devnet-deploy.yml + mainnet-deploy.yml — balance check floor 5 → 20 SOL (1.6× empirical for priority-fee spikes + multi-program buffer-account temporaries during deploy).
Comment in each step's header documents the per-program breakdown so future devs don't have to re-derive empirically.
docs/operations/cd-pipeline.md — all 5 refs to "5 SOL" updated to "20 SOL" with cost rationale inline.

Recovery note for operator

Rehearsal-1f's failed deploy left a buffer account AqUsc5qqMG7bk5W6kZZzbhsfbWeuEMbEzuVpLkwr4PSH with escrowed lamports. Recoverable via:

solana program close AqUsc5qqMG7bk5W6kZZzbhsfbWeuEMbEzuVpLkwr4PSH \
  --recipient 5ZpFtJePb2hGKhG9RJ6Fdwmo5y8wuwKXZZcKttoN1Jgo \
  --url devnet

Optional — devnet SOL is free, faster to just request more airdrops. Important for mainnet: the same lock-up pattern would happen there too if balance is insufficient. Mainnet floor at 20 SOL is critical.

Rehearsal chronology (full saga)

Rehearsal	PR	Surfaced bug
1a	(pre-secret-setup)	secret restore failed (expected; no secret yet)
1b	#389	`solana balance --output json \| jq` returns null
1c	#390	`anchor build` (no --no-idl) hits IDL gen error
1d	#391	`anchor deploy` ignores ANCHOR_WALLET env
1e	#392	`anchor deploy` ignores `solana config`, reads Anchor.toml
1f	this PR	Real cost is 12.5 SOL, not 3-4 SOL
1g (next)	—	Should actually complete successfully

After this merges + operator funds wallet to 25+ SOL: rehearsal-1g either succeeds or the next failure is RPC/Solana-side, not workflow.

Test plan

Lint green
CI all standard lanes stay green
Empirical: operator pushes devnet-deploy-v$(date +%Y%m%d)-rehearsal-1g after wallet topped to 25+ SOL — should complete anchor deploy successfully and upload artifact

Freeze status

Permitted (SEV-046 follow-up). 3 files, ~37 lines net.

Recommended merge method

Squash.

https://claude.ai/code/session_01YapZy1Z5gzbV5EammBkSQm

Generated by Claude Code

Rehearsal-1f finally got past every workflow bug and surfaced the real cost: anchor deploy of 4 programs needs ~12.5 SOL, not the ~3-4 SOL my original docs claimed. Run failed mid-deploy with: Error: Account 5ZpFtJePb2hGKhG9RJ6Fdwmo5y8wuwKXZZcKttoN1Jgo has insufficient funds for spend (6.13730712 SOL) + fee (0.00438 SOL) Per-program rent-exempt minimum is 2× bytecode size: roundfi_core ~880KB → ~6.14 SOL roundfi_reputation ~370KB → ~2.6 SOL roundfi_yield_kamino ~294KB → ~2.0 SOL roundfi_yield_mock ~254KB → ~1.8 SOL TOTAL ≈ 12.5 SOL + ~0.05 tx fees Fixes: - devnet-deploy.yml + mainnet-deploy.yml balance floor 5 → 20 SOL (20 = 1.6× empirical for priority-fee spikes + temporary buffer accounts during multi-program deploys) - Header comment updated with per-program breakdown so future devs don't have to re-derive empirically. - docs/operations/cd-pipeline.md: all 5 refs to "5 SOL" updated to "20 SOL" with the cost rationale inline. This is the LAST plausible workflow change for SEV-046 execution. Next failure mode (if any) is a Solana-side RPC issue, not a workflow bug — the workflow itself is now demonstrably correct through `anchor deploy` invocation. Note for operator on rehearsal-1f's burned SOL: the buffer account AqUsc5qqMG7bk5W6kZZzbhsfbWeuEMbEzuVpLkwr4PSH has the lamports escrowed. Recoverable via: solana program close AqUsc5qqMG7bk5W6kZZzbhsfbWeuEMbEzuVpLkwr4PSH \ --recipient 5ZpFtJePb2hGKhG9RJ6Fdwmo5y8wuwKXZZcKttoN1Jgo \ --url devnet Optional — devnet SOL is free; faster to just airdrop more. https://claude.ai/code/session_01YapZy1Z5gzbV5EammBkSQm

Companion to docs/operations/cd-pipeline.md (architecture spec). This is the empirical lessons-learned counterpart documenting why SEV-046's "clean" PR #388 took 5 follow-up PRs (#389-#393) before a single anchor deploy actually completed. Contents: - Headline: 5 bugs, each one-line, none catchable by lint/typecheck - Per-rehearsal chronology (1a → 1g) - Per-bug root cause analysis (JSON parse trap, anchor-syn IDL, anchor wallet resolution chain, empirical SOL cost) - What would have prevented it: localnet smoke-test workflow that exercises the full deploy path against solana-test-validator - Mainnet-deploy mitigation summary (all 5 fixes already in) - Outstanding operator-side work tracker Filed as follow-up: the localnet smoke-test workflow design is sketched but not implemented in scope of SEV-046. Tracking for post-canary cleanup. Generalized lesson burned in: workflow code is untestable except by running it. The CI lanes pass; the workflow still doesn't work. https://claude.ai/code/session_01YapZy1Z5gzbV5EammBkSQm

#395) Documents the 5-PR chain (#389-#393) that the SEV-046 CD scaffolding (PR #388) needed before a single anchor deploy actually ran end-to-end. Companion to docs/operations/cd-pipeline.md (architecture spec) — this is the empirical lessons-learned counterpart. Generalized lesson: workflow code is untestable except by running it. CI lanes can be green while the workflow itself is broken in ways that only surface against real runner conditions. https://claude.ai/code/session_01YapZy1Z5gzbV5EammBkSQm

…closed First green end-to-end CD devnet deploy after the 5-PR bug chain (#389–#393). Run 26086314957 deployed all 4 RoundFi programs on a clean ubuntu-latest runner, artifact captured, keypair scrubbed. Changes: - docs/operations/rehearsal-logs/2026-05-19-SEV-046-rehearsal-1g-success.md — full run record with program IDs, Solscan links, cost breakdown, next-rehearsal plan. - docs/operations/rehearsal-logs/2026-05-18-SEV-046-rehearsal-saga.md — row 1g flipped from "pending" to ✓ green + headline update note. - docs/operations/mainnet-canary-plan.md §3.3 — CD pipeline checkbox flipped [x] with link to the 1g success log. Stretch goal of 3× clean still at 1/3. - docs/security/internal-audit-findings.md — SEV-046 tracker row updated with PR chain breadcrumb (#388 + #389–#393 + #395) + empirical close status. - CHANGELOG.md — new [Unreleased] entry summarizing the rehearsal 1g outcome above the existing #272 scaffolding entry. §3.3 strict "at least once" criterion: ✓ satisfied. 3× clean reproducibility (cd-pipeline.md §"Rehearsal protocol"): 1/3. Operator can now disparar rehearsals 2 + 3 with no expected code changes. https://claude.ai/code/session_01YapZy1Z5gzbV5EammBkSQm

…tag (#398) SEV-046 rehearsal 2 abort log + 1g tag correction. Operator disparou rehearsal-2 post-1g merge; workflow correctly aborted at the 20 SOL balance gate (deployer wallet was at 12.38 SOL post-1g — 1g consumed the ~12.5 SOL PR #393 predicted). Floor working as designed, not a code bug. Recorded across saga doc + 1g success doc + canary-plan §3.3 checkbox. Side fix: 1g doc Tag field corrected from 20260518 → 20260519 to match actual git tag. §3.3 strict "at least once" criterion remains satisfied by 1g; 3× stretch goal still at 1/3, NOT a mainnet blocker.

Second green CD devnet deploy via run 26115425088 — 4 programs deployed with fresh keypairs, 12.62 SOL consumed (within 1% of PR #393 prediction). Recorded full run + 1g-vs-2b reproducibility table (8/8 properties match or within 1%) + saga continuation row + canary plan §3.3 checkbox 1/3 → 2/3 + CHANGELOG. Reproducibility empirically demonstrated: the 5-PR bug chain in 1b–1f (PRs #389–#393) has no siblings. Remaining gap is operator-side SOL availability for rehearsal-3 (post-2b balance is 9.76 SOL; needs top-up to clear 20 SOL floor). §3.3 strict criterion ("at least once") remains satisfied since 1g. 3× stretch goal is NOT a mainnet blocker.

alrimarleskovar merged commit e58ca1c into main May 18, 2026
8 checks passed

alrimarleskovar mentioned this pull request May 18, 2026

docs(SEV-046): rehearsal saga log — 5-PR chain to first working deploy #395

Merged

2 tasks

This was referenced May 19, 2026

docs(canary-plan sync): SEV-046 rehearsal 1g + Pass-13 alignment sweep (#230 / #233 / #249) #396

Merged

docs(SEV-046): log rehearsal-2 abort (balance gate working) + fix 1g tag #398

Merged

alrimarleskovar mentioned this pull request May 19, 2026

docs(SEV-046): rehearsal 2b green — 3× stretch goal now 2/3 #399

Merged

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(SEV-046): bump deploy SOL floor 5→20 (empirical 12.5 SOL needed)#393

fix(SEV-046): bump deploy SOL floor 5→20 (empirical 12.5 SOL needed)#393
alrimarleskovar merged 1 commit into
mainfrom
claude/implement-roundfi-desktop-SRV6l

alrimarleskovar commented May 18, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

alrimarleskovar commented May 18, 2026

TL;DR

What rehearsal-1f surfaced

Empirical per-program cost

Fix

Recovery note for operator

Rehearsal chronology (full saga)

Test plan

Freeze status

Recommended merge method

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants