* Fix missing block number issue on forced canonicalization
There is this issue about missing block numbers on forced canonicalization. I looked over the code
now 10000 times and there are possible ways this can be triggered, but I don't really know how this
is triggered. So, this pr is going to solve the symptom and not the cause. The block number to hash
mapping is set when we import a new best block. Forced canonicalization will now stop at the best
block and it will canonicalize the other blocks later when the best block moved. As the error
reports indicated that this issue mainly happened on major sync, there should not be any forks, so
not doing the canonicalization directly shouldn't be that harmful. All known implementations should
import all blocks as best block on major sync anyway (I mean somewhere there is the bug, but I
didn't yet found it).
I will also do some changes to Cumulus around some potential culprit for this issue.
Closes: https://github.com/paritytech/substrate/issues/12613
* Add some docs
* Fix fix
* Review comments
* Review comments
* jsonrpsee v0.16
add backwards compatibility
run old http server on http only
* cargo fmt
* update jsonrpsee 0.16.1
* less verbose cors log
* fix nit in log: WS -> HTTP
* revert needless changes in Cargo.lock
* remove unused features in tower
* fix nits; add client-core feature
* jsonrpsee v0.16.2
* Ensure that we inform all tasks to stop before starting the 60 seconds shutdown
The change of waiting in maximum 60 seconds for the node to shutdown actually introduced a bug. We
were actually waiting always 60 seconds as we didn't informed our tasks to shutdown. The solution to
this problem is to drop the task manager as this will then inform all tasks to end. It also adds
tests to ensure that the behaviors work as expected. (This should already have been done in the
first pr! :()
* ".git/.scripts/fmt.sh" 1
Co-authored-by: command-bot <>
* Move import queue out of `sc-network`
Add supplementary asynchronous API for the import queue which means
it can be run as an independent task and communicated with through
the `ImportQueueService`.
This commit removes removes block and justification imports from
`sc-network` and provides `ChainSync` with a handle to import queue so
it can import blocks and justifications. Polling of the import queue is
moved complete out of `sc-network` and `sc_consensus::Link` is
implemented for `ChainSyncInterfaceHandled` so the import queue
can still influence the syncing process.
* Fix tests
* Apply review comments
* Apply suggestions from code review
Co-authored-by: Bastian Köcher <git@kchr.de>
* Update client/network/sync/src/lib.rs
Co-authored-by: Bastian Köcher <git@kchr.de>
Co-authored-by: Bastian Köcher <git@kchr.de>
* Don't indefinitely on shutting down Tokio
Now we wait in maximum 60 seconds before we shutdown the node. Tasks are may be leaked and leading
to some data corruption.
* Drink less :thinking_face:
* cli: Improve pruning documentation
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Keep `finalized` notation and remove `canonical` one
* cli: Fix cargo doc
* cli: `PruningModeClap` IR enum
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Convert PruningModeClap into pruning modes
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Use `PruningModeClap`
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Rename to `DatabasePruningMode`
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Implement `FromStr` instead of `clap::ValueEnum`
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* Update client/cli/src/params/pruning_params.rs
Co-authored-by: Bastian Köcher <git@kchr.de>
* Fix clippy
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* cli: Add option documentation back
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
* Apply suggestions from code review
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
Co-authored-by: Bastian Köcher <git@kchr.de>
* implement crate publishing from CI
* fix indentation
* use resource_group for job exclusivity
ensure that at most one instance of the publish-crates job is running at any given time to prevent race conditions
* correct publish = false
* Remove YAML anchors as GitLab's `extends:` doesn't need it
* Temporarily force cache upload for the new jobs
* Revert `RUSTY_CACHIER_FORCE_UPLOAD`
* pin libp2p-tcp=0.37.0 for sc-telemetry
* Revert "pin libp2p-tcp=0.37.0 for sc-telemetry"
This reverts commit 29146bfad6c31e8cf0e2f17ad92a71bb81a373af.
* always collect generated crates
* increase timeout for publish-crates-template
* Force upload the new job cache again
* Revert "Force upload the new job cache again"
This reverts commit 5a5feee1b2c51fdef768b25a76be4c3949ec1c99.
* reformat
* improve timeout explanation
* s/usual/average
Co-authored-by: Vladimir Istyufeev <vladimir@parity.io>
client/mmr: persisting gadget state across runs
Fixes#12780
* client/mmr: on init do canonicalization catch-up
* client/mmr: add more tests
* client/mmr: persist gadget progress in aux db
* client/mmr: add more tests
* client/mmr: replace async_std with tokio
* remove leftover comment
* address review comments
Signed-off-by: acatangiu <adrian@parity.io>
Introduce bounds on the justifications and votes queues, so they do not grow forever if voter cannot make progress and consume from them. When bounds are hit, new votes or justifications get dropped.
* use a BTreeMap and check for bounds
* cargo fmt
* use usize
Co-authored-by: Adrian Catangiu <adrian@parity.io>
* Replace deprecated libp2p feature specs with correct ones
* Bump tokio to 1.21.2
* Replace async-std libp2p primitives with tokio ones
* minor: rustfmt
* Fix TestNet to run initialization in the tokio context
* Convert telemetry test from async-std to tokio
* Convert notifications tests from async-std to tokio
* Convert chain sync tests from async-std to tokio
* Ditch async-std completely
* Make executor mandatory
* Bump tokio to 1.22.0
* minor: rustfmt
* Explicitly use tokio runtime in tests
* Move more tests to explicit tokio runtime
* Explicitly set multithreaded runtime in tokio test
* minor: rustfmt
* minor: fix comment
* Replace async-std with tokio in MMR tests
* Move create_inherent_data call to use side
* Make provide_inherent_data async
* Fix tests
* Apply suggestions from code review
Co-authored-by: Bastian Köcher <git@kchr.de>
* Log errors
* Fix test
* Fix test
* fix
* Deduplicate test code
* fix
* flag
* Update client/consensus/slots/src/lib.rs
Co-authored-by: Bastian Köcher <git@kchr.de>
* Revert "Deduplicate test code"
This reverts commit ba46adbe089329c78cd69ccdb08e27ed67bd77cf.
* Fix test
* remove commented out code
* minor to start CI run
* start CI
* Update client/consensus/slots/src/lib.rs
Co-authored-by: Marcin S. <marcin@bytedude.com>
* Apply PR suggestions
* Apply PR suggestions
* Update client/consensus/slots/src/lib.rs
Co-authored-by: Bastian Köcher <git@kchr.de>
* minor
* kickoff CI
* PR suggestions
* Compute remaining duration instead of using slot_info.duration
* Don't rely on sub implementation for Instant
* Apply PR suggestions
* Use saturating_duration_since
Co-authored-by: Bastian Köcher <git@kchr.de>
Co-authored-by: Marcin S. <marcin@bytedude.com>
Co-authored-by: parity-processbot <>
* client/beefy: fix on-demand justif sync for old blocks
When receiving BEEFY justifications for old blocks the state might
be pruned for them, in which case justification verification fails
because BEEFY validator set cannot be retrieved from runtime state.
Fix this by having the voter give the validator set to the
`OnDemandJustificationsEngine` as request information. On receiving
a BEEFY justification for requested block, the provided validator
set will be used to validate the justification.
Signed-off-by: acatangiu <adrian@parity.io>
* Apply suggestions from code review
Co-authored-by: Bastian Köcher <git@kchr.de>
* impl review suggestions
* client/beefy: fail initialization if state unavailable
* beefy: remove spammy log
Signed-off-by: acatangiu <adrian@parity.io>
Co-authored-by: parity-processbot <>
Co-authored-by: Bastian Köcher <git@kchr.de>
* ed25519_verify: Support using dalek for historical blocks
The switch from `ed25519-dalek` to `ed25519-zebra` was actually a breaking change. `ed25519-zebra`
is more permissive. To support historical blocks when syncing a chain this pull request introduces
an externalities extension `UseDalekExt`. This extension is just used as a signaling mechanism to
`ed25519_verify` to use `ed25519-dalek` when it is present. Together with `ExtensionBeforeBlock` it
can be used to setup a node in way to sync historical blocks that require `ed25519-dalek`, because
they included a transaction that verified differently as when using `ed25519-zebra`.
This feature can be enabled in the following way. In the chain service file, directly after the
client is created, the following code should be added:
```
use sc_client_api::ExecutorProvider;
client.execution_extensions().set_extensions_factory(
sc_client_api::execution_extensions::ExtensionBeforeBlock::<Block, sp_io::UseDalekExt>::new(BLOCK_NUMBER_UNTIL_DALEK_SHOULD_BE_USED)
);
```
* Fix doc
* More fixes
* Update client/api/src/execution_extensions.rs
Co-authored-by: André Silva <123550+andresilva@users.noreply.github.com>
* Fix merge and warning
* Fix docs
Co-authored-by: André Silva <123550+andresilva@users.noreply.github.com>
* client/beefy: remove high-freq network events from main loop
Network events are many and very frequent, remove the net-event-stream
from the main voter loop and drastically reduce BEEFY voter task
'wakeups'.
Instead have the `GossipValidator` track known peers as it already
has callbacks for that coming from `GossipEngine`.
Signed-off-by: acatangiu <adrian@parity.io>
* client/beefy: prepare worker for persisting state
* client/beefy: persist voter state
* client/beefy: initialize persistent state
* client/beefy: try to vote from the very beginning
Now that voter is initialized from persistent state, it makes
sense that it can attempt voting right away. This also helps
the genesis case when we consider block `One` as mandatory.
* client/beefy: add tests for voter state db
* client/beefy: persist voter state as soon as initialized
* client/beefy: make sure min-block-delta is at least 1
* client/beefy: persist state after voting
Persist state after handling self vote to avoid double voting in case
of voter restarts.
* client/beefy: persist state after handling mandatory block vote
For mandatory blocks we want to make sure we're not losing votes
in case of crashes or restarts, since voter will not make further
progress without finalizing them.
* frame/beefy: use GENESIS_AUTHORITY_SET_ID on pallet genesis
* client/beefy: initialize voter at either genesis or last finalized
To guarantee unbroken chain of mandatory blocks justifications, voter
will always resume from either last BEEFY-justified block or
`pallet-beefy` genesis, whichever is more recent.
Initialization walks back the chain from latest GRANDPA finalized
block looking for one of the above. Along the way, it also records
and enqueues for processing any BEEFY mandatory blocks that have
been already GRANDPA finalized but not BEEFY finalized.
* client/beefy: decouple voter init from aux db state load
* client/beefy: fix voter init tests
* remove debug prints
* gadget future must be type ()
* fix init from last justification
Signed-off-by: Adrian Catangiu <adrian@parity.io>
Before it was using `build_storage` and `assimilate_storage` was returning an error. However, there
was no real reason for `assimilate_storage` to return an error. This pr implements
`assimilate_storage` and uses the default `build_storage` of the trait.
* Remove `sp_tasks::spawn` API and related code
* Remove `RuntimeTasks::{spawn, join}` host functions
* remove unused
* Remove a few more tests that I forgot to remove
Co-authored-by: Shawn Tabrizi <shawntabrizi@gmail.com>
* histor. batch proof: make best block arg optional
* correct testing range
* make generate_batch_proof stub for historical
* merge generate_{historical_}batch_proof functions
* merge generate_{batch_}proof functions
* merge verify_{batch_}proof functions
* merge verify_{batch_}proof_stateless functions
* remove {Leaf}Proof
Not utilized by API anymore, so superfluous.
Removal consistent with prior changes to just use "batch" proof API.
* rename BatchProof->Proof
no need to qualify if only one universal proof type.
* cleanup
* expose verify_proof rpc api
* document verify_proof
* expose verify_proof_stateless rpc api
* add optional BlockHash to mmr_root rpc api
* fixup! expose verify_proof rpc api
* fix documentation phrasing
Co-authored-by: Adrian Catangiu <adrian@parity.io>
* documentation grammar
Co-authored-by: Adrian Catangiu <adrian@parity.io>
* define mmr error msgs together with error enum
Co-authored-by: Serban Iorga <serban@parity.io>
* fixup! define mmr error msgs together with error enum
* map decoding errors to CallError::InvalidParams
Co-authored-by: Serban Iorga <serban@parity.io>
* fixup! map decoding errors to CallError::InvalidParams
Co-authored-by: Adrian Catangiu <adrian@parity.io>
Co-authored-by: parity-processbot <>
Co-authored-by: Serban Iorga <serban@parity.io>
* Remove all not required nodes from the epoch-changes tree
Some outdated nodes were left there because of the predicate
* Test to excercise the fix
* Add a fork on genesis to the test
* Fix typo in comments
* Prune discarded blocks immediately
* state-db refactoring part 1
* Some renames
* Get rid of pending state
* Revert "Prune discarded blocks immediately"
This reverts commit b60d98c569e8af18d99087da93f0911d4f24006e.
* Cleanup
* Make clippy happy
* Minor changes
* Do not update peer information if ancestor search is in progress
If block announcement is received from a peer while ancestor search
for that same peer is still in progress, do not update the peer's best
hash and best number as that causes the ancestor search to yield
different information from what was expected and can cause, for example,
a fork of lower height not be be downloaded.
* Block until peers are in sync
Bastian Köcher