Remove StateTrieMigration pallet - V0→V1 migration complete #1040

nkpar · 2025-12-18T16:16:59Z

Summary

The state trie migration from V0 to V1 format has been completed on all Polkadot ecosystem chains:

Chain	Top Items Migrated	Child Items	Status
Polkadot Relay	2,192,421	10,423	✅ Complete
Asset Hub Polkadot	5,448,606	0	✅ Complete
Asset Hub Kusama	288,016	0	✅ Complete

This PR removes the pallet-state-trie-migration from these runtimes and adds RemovePallet migrations to clean up the storage.

Changes

For each runtime:

Removed pallet-state-trie-migration dependency from Cargo.toml
Removed pallet configuration (impl pallet_state_trie_migration::Config)
Removed from construct_runtime! macro
Added RemovePallet<StateTrieMigrationPalletName, DbWeight> migration
Removed associated MigController accounts and tests (Asset Hubs only)

Reference

Follows the same pattern as PR Remove state trie migration from kusama, add it to polkadot #170 (Kusama relay chain removal)
Tracking issue: State migration #74

Test Plan

CI passes
RemovePallet migration tested via try-runtime (if enabled)

Closes #74 (for Polkadot ecosystem chains)

Update value of MigrationMaxKeyLen to proceed with migration via bot on both KAH and PAH

Set ControlOrigin to MigController

Migration limit could be set either by fellowship or MigController

The state trie migration from V0 to V1 format has been completed on: - Polkadot relay chain (2,192,421 top items + 10,423 child items) - Asset Hub Polkadot (5,448,606 top items) - Asset Hub Kusama (288,016 top items) This PR removes the pallet-state-trie-migration from these runtimes and adds RemovePallet migrations to clean up the storage. Closes polkadot-fellows#74 (for Polkadot ecosystem chains)

github-actions · 2025-12-18T21:25:39Z

Review required! Latest push from author must always be reviewed

github-actions

Found 3 runtime-safety issues: (1) unreachable!() introduced in runtime conversion paths (panic risk), (2) potential SCALE decoding / runtime failure risk if any StateTrieMigration hold reasons still exist in on-chain Balances::Holds, and (3) minor CHANGELOG formatting noise (blank lines/whitespace).

Suggestions that couldn't be attached to a specific line

relay/polkadot/src/ah_migration/phase1.rs:137, 156-157

Removal of the StateTrieMigration match arms from call filtering and (more importantly) from RuntimeHoldReason portable conversion increases the risk of a runtime-breaking upgrade if any on-chain Balances::Holds entries (or other storage/XCM payloads) still encode the StateTrieMigration hold reason. RemovePallet only clears the pallet's own storage prefix; it does not clear holds stored under pallet_balances. Actionable mitigation: (a) add/confirm a dedicated runtime migration that scans pallet_balances::Holds and removes any StateTrieMigration holds (or asserts none exist in try-runtime pre/post upgrade), and/or (b) keep the enum variant for at least one release after removing the pallet to maintain decoding compatibility.

github-actions · 2026-01-08T12:31:30Z

system-parachains/asset-hubs/asset-hub-kusama/src/ah_migration/mod.rs

+			PortableHoldReason::StateTrieMigration(_) =>
+				unreachable!("StateTrieMigration pallet removed - no holds should exist"),


PortableHoldReason::StateTrieMigration(_) => unreachable!(...) will panic if that variant is ever encountered. Even if "should never happen", a panic in runtime code is consensus-critical (can abort block execution). Replace this with non-panicking handling: e.g., change the conversion to TryFrom returning Result<RuntimeHoldReason, Error> and make callers handle/propagate the error, or (safer for backwards compatibility) keep a RuntimeHoldReason::StateTrieMigration variant (even if the pallet is removed) so old data/messages can still decode without panicking.

github-actions · 2026-01-08T12:31:30Z

system-parachains/asset-hubs/asset-hub-polkadot/src/ah_migration/mod.rs

+			PortableHoldReason::StateTrieMigration(_) =>
+				unreachable!("StateTrieMigration pallet removed - no holds should exist"),


Same issue as Asset Hub Kusama: unreachable!() in a runtime conversion is a panic path. If any stale/malicious/old-format data causes PortableHoldReason::StateTrieMigration to be processed, the runtime can panic. Prefer a non-panicking conversion (TryFrom) or preserve a compatible enum variant and map it safely.

github-actions · 2026-01-08T12:31:30Z

CHANGELOG.md

+### Removed
+
+- Remove StateTrieMigration pallet from Polkadot relay chain, Asset Hub Polkadot, and Asset Hub Kusama - V0→V1 state trie migration complete ([polkadot-fellows/runtimes#1040](https://github.com/polkadot-fellows/runtimes/pull/1040))
+


The new entry under ### Removed includes extra blank lines (and appears to introduce whitespace-only lines). This tends to trip markdown formatting/linting and makes the changelog noisier. Remove the empty lines and ensure the bullet is directly under the heading (no whitespace-only lines).

acatangiu · 2026-01-08T12:28:43Z

relay/polkadot/src/lib.rs

+			StateTrieMigrationPalletName,
+			<Runtime as frame_system::Config>::DbWeight,
+		>,
+		KickOffAhm<Runtime>,


can you also remove this one (KickOffAhm)?

acatangiu · 2026-01-08T12:31:03Z

system-parachains/asset-hubs/asset-hub-kusama/src/ah_migration/mod.rs

+			PortableHoldReason::StateTrieMigration(_) =>
+				unreachable!("StateTrieMigration pallet removed - no holds should exist"),


why is there still an enum variant for StateTrieMigration without the pallet? can we remove completely?

and even if not, please remove the unreachable!() - it's a footgun if we later add back the pallet and miss updating this

What prevents us from removing the entire ah_migration.mod? Then we do not have to care about this issue.

why is there still an enum variant for StateTrieMigration without the pallet?

@nkpar we need to clean it up here:

runtimes/pallets/rc-migrator/src/types.rs

Line 654 in b31ce90

StateTrieMigration(pallet_state_trie_migration::HoldReason),

not sure how this enum is used, but better add there indexes:

pub enum PortableHoldReason { + #[codec(index = 0)] Preimage(pallet_preimage::HoldReason), + #[codec(index = 1)] Staking(pallet_staking::HoldReason), - #[codec(index = 2)] - StateTrieMigration(pallet_state_trie_migration::HoldReason), + #[codec(index = 3)] DelegatedStaking(pallet_delegated_staking::HoldReason), + #[codec(index = 4)] Session(pallet_session::HoldReason), + #[codec(index = 5)] XcmPallet(pallet_xcm::HoldReason), }

What prevents us from removing the entire ah_migration.mod? Then we do not have to care about this issue.

@clangenb Partially, it is still used: #1036
and there is a big PR removing all the stuff: #1016

acatangiu · 2026-01-08T12:32:27Z

system-parachains/asset-hubs/asset-hub-polkadot/src/ah_migration/mod.rs

-			PortableHoldReason::StateTrieMigration(state_trie_migration) =>
-				StateTrieMigration(state_trie_migration),
+			PortableHoldReason::StateTrieMigration(_) =>
+				unreachable!("StateTrieMigration pallet removed - no holds should exist"),


same comment as for AHK, don't like the unreachable

bkontur · 2026-01-08T13:29:26Z

Cargo.lock

 "pallet-staking-async",
 "pallet-staking-async-rc-client",
 "pallet-staking-runtime-api",
- "pallet-state-trie-migration",


@nkpar I would also remove pallet-state-trie-migration = { workspace = true } from:

runtimes/pallets/ah-migrator/Cargo.toml

Line 39 in b31ce90

pallet-state-trie-migration = { workspace = true }

runtimes/pallets/rc-migrator/Cargo.toml

Line 47 in b31ce90

pallet-state-trie-migration = { workspace = true }

and from the main Cargo.toml:

runtimes/Cargo.toml

Line 174 in b31ce90

pallet-state-trie-migration = { version = "49.0.0", default-features = false }

kianenigma

Existing comments should be addressed, otherwise a great cleanup to, wishing to see more of such PRs to remove legacy stuff.

nkpar and others added 6 commits April 8, 2025 16:08

Update value of MigrationMaxKeyLen

4d3bf07

Update value of MigrationMaxKeyLen to proceed with migration via bot on both KAH and PAH

Update ControlOrigin

89bdacb

Set ControlOrigin to MigController

Update ControlOrigin for PAH

ec4f19c

have an option to set a migration lim

c7f7859

Migration limit could be set either by fellowship or MigController

rust fmt

7e81390

nkpar force-pushed the remove-state-trie-migration branch from 1a650d3 to 0d02e8e Compare December 18, 2025 16:17

nkpar mentioned this pull request Dec 18, 2025

[Westend] Remove StateTrieMigration pallet - V0→V1 migration complete paritytech/polkadot-sdk#10694

Open

2 tasks

bkchr approved these changes Dec 18, 2025

View reviewed changes

Merge upstream main and re-apply StateTrieMigration removal

04faa90

github-actions bot requested a review from bkchr December 18, 2025 21:25

nkpar and others added 6 commits December 18, 2025 21:41

Fix StateTrieMigration references in AH migration code

cd502ef

Fix rustfmt comment alignment

f0a5a7b

Merge branch 'main' into remove-state-trie-migration

b3bfe01

chore: retrigger CI

a13f0db

Merge branch 'main' into remove-state-trie-migration

0037c60

Merge branch 'main' into remove-state-trie-migration

39a57e6

github-actions bot reviewed Jan 8, 2026

View reviewed changes

acatangiu reviewed Jan 8, 2026

View reviewed changes

bkontur reviewed Jan 8, 2026

View reviewed changes

kianenigma approved these changes Jan 9, 2026

View reviewed changes

		PortableHoldReason::StateTrieMigration(_) =>
		unreachable!("StateTrieMigration pallet removed - no holds should exist"),

		### Removed

		- Remove StateTrieMigration pallet from Polkadot relay chain, Asset Hub Polkadot, and Asset Hub Kusama - V0→V1 state trie migration complete ([polkadot-fellows/runtimes#1040](https://github.com/polkadot-fellows/runtimes/pull/1040))

Remove StateTrieMigration pallet - V0→V1 migration complete #1040

Are you sure you want to change the base?

Remove StateTrieMigration pallet - V0→V1 migration complete #1040

Uh oh!

Conversation

nkpar commented Dec 18, 2025

Summary

Changes

Reference

Test Plan

Uh oh!

github-actions bot commented Dec 18, 2025

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Suggestions that couldn't be attached to a specific line

Uh oh!

github-actions bot Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

acatangiu Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

acatangiu Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

clangenb Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

bkontur Jan 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bkontur Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

acatangiu Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

bkontur Jan 8, 2026

Choose a reason for hiding this comment

Uh oh!

kianenigma left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

bkontur Jan 8, 2026 •

edited

Loading