When a peer is stopped due to some network issue, the Reactor calls scheduler#handleRemovePeer, which removes the peer from the scheduler. BUT the peer stays in the processor, which sometimes could lead to "duplicate block enqueued by processor" panic WHEN the same block is requested by the scheduler again from a different peer. The solution is to return scPeerError, which will be propagated to the processor. The processor will clean up the blocks associated with the peer in purgePeer.
Closes#5513, #5517
## Description
This PR wraps the stdlib sync.(RW)Mutex & godeadlock.(RW)Mutex. This enables using go-deadlock via a build flag instead of using sed to replace sync with godeadlock in all files
Closes: #3242
- drop Height & Base from StatusRequest
It does not make sense nor it's used anywhere currently. Also, there
seem to be no trace of these fields in the ADR-40 (blockchain reactor
v2).
- change PacketMsg#EOF type from int32 to bool
Integrates the blockchain v2 reactor with state sync, fixes#4765. This mostly involves deferring fast syncing until after state sync completes. I tried a few different approaches, this was the least effort:
* `Reactor.events` is `nil` if no fast sync is in progress, in which case events are not dispatched - most importantly `AddPeer`.
* Accept status messages from unknown peers in the scheduler and register them as ready. On fast sync startup, broadcast status requests to all existing peers.
* When switching from state sync, first send a `bcResetState` message to the processor and scheduler to update their states - most importantly the initial block height.
* When fast sync completes, shut down event loop, scheduler and processor, and set `events` channel to `nil`.
The event loop uses a `select` on multiple channels. However, reading from a closed channel in Go always yields the channel's zero value. The processor and scheduler close their channels when done, and since these channels are always ready to receive, the event loop keeps spinning on them.
This changes `routine.terminate()` to not close the channel, and also removes `stopDemux` and instead uses `events` channel closure to signal event loop termination.
Fixes#4687.
Fixes#828. Adds state sync, as outlined in [ADR-053](https://github.com/tendermint/tendermint/blob/master/docs/architecture/adr-053-state-sync-prototype.md). See related PRs in Cosmos SDK (https://github.com/cosmos/cosmos-sdk/pull/5803) and Gaia (https://github.com/cosmos/gaia/pull/327).
This is split out of the previous PR #4645, and branched off of the ABCI interface in #4704.
* Adds a new P2P reactor which exchanges snapshots with peers, and bootstraps an empty local node from remote snapshots when requested.
* Adds a new configuration section `[statesync]` that enables state sync and configures the light client. Also enables `statesync:info` logging by default.
* Integrates state sync into node startup. Does not support the v2 blockchain reactor, since it needs some reorganization to defer startup.
to prevent malicious nodes from sending us large messages (~21MB, which
is the default `RecvMessageCapacity`)
This allows us to remove unnecessary `maxMsgSize` check in `decodeMsg`. Since each channel has a msg capacity set to `maxMsgSize`, there's no need to check it again in `decodeMsg`.
Closes#1503
<!-- < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < ☺
v ✰ Thanks for creating a PR! ✰
v Before smashing the submit button please review the checkboxes.
v If a checkbox is n/a - please still include it but + a little note why
☺ > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -->
Fixes an issue reported in https://github.com/tendermint/tendermint/issues/4595#issuecomment-612667441.
Not sure if this is sufficient to fully remove the reactor, but it fixes the immediate problem.
______
For contributor use:
- [x] Wrote tests
- [x] ~Updated CHANGELOG_PENDING.md~
- [x] Linked to Github issue with discussion and accepted design OR link to spec that describes this work.
- [x] ~Updated relevant documentation (`docs/`) and code comments~
- [x] Re-reviewed `Files changed` in the Github PR explorer
* Added BlockStore.DeleteBlock()
* Added initial block pruner prototype
* wip
* Added BlockStore.PruneBlocks()
* Added consensus setting for block pruning
* Added BlockStore base
* Error on replay if base does not have blocks
* Handle missing blocks when sending VoteSetMaj23Message
* Error message tweak
* Properly update blockstore state
* Error message fix again
* blockchain: ignore peer missing blocks
* Added FIXME
* Added test for block replay with truncated history
* Handle peer base in blockchain reactor
* Improved replay error handling
* Added tests for Store.PruneBlocks()
* Fix non-RPC handling of truncated block history
* Panic on missing block meta in needProofBlock()
* Updated changelog
* Handle truncated block history in RPC layer
* Added info about earliest block in /status RPC
* Reorder height and base in blockchain reactor messages
* Updated changelog
* Fix tests
* Appease linter
* Minor review fixes
* Non-empty BlockStores should always have base > 0
* Update code to assume base > 0 invariant
* Added blockstore tests for pruning to 0
* Make sure we don't prune below the current base
* Added BlockStore.Size()
* config: added retain_blocks recommendations
* Update v1 blockchain reactor to handle blockstore base
* Added state database pruning
* Propagate errors on missing validator sets
* Comment tweaks
* Improved error message
Co-Authored-By: Anton Kaliaev <anton.kalyaev@gmail.com>
* use ABCI field ResponseCommit.retain_height instead of retain-blocks config option
* remove State.RetainHeight, return value instead
* fix minor issues
* rename pruneHeights() to pruneBlocks()
* noop to fix GitHub borkage
Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>
* format: add format cmd & goimport repo
- replaced format command
- added goimports to format command
- ran goimports
Signed-off-by: Marko Baricevic <marbar3778@yahoo.com>
* fix outliers & undo proto file changes
The work includes the reactor which ties together all the seperate routines involved in the design of the blockchain v2 refactor. This PR replaces #4067 which got far too large and messy after a failed attempt to rebase.
## Commits:
* Blockchainv 2 reactor:
+ I cleaner copy of the work done in #4067 which fell too far behind and was a nightmare to rebase.
+ The work includes the reactor which ties together all the seperate routines involved in the design of the blockchain v2 refactor.
* fixes after merge
* reorder iIO interface methodset
* change iO -> IO
* panic before send nil block
* rename switchToConsensus -> trySwitchToConsensus
* rename tdState -> tmState
* Update blockchain/v2/reactor.go
Co-Authored-By: Bot from GolangCI <42910462+golangcibot@users.noreply.github.com>
* remove peer when it sends a block unsolicited
* check for not ready in markReceived
* fix error
* fix the pcFinished event
* typo fix
* add documentation for processor fields
* simplify time.Since
* try and make the linter happy
* some doc updates
* fix channel diagram
* Update adr-043-blockchain-riri-org.md
* panic on nil switch
* liting fixes
* account for nil block in bBlockResponseMessage
* panic on duplicate block enqueued by processor
* linting
* goimport reactor_test.go
Co-authored-by: Bot from GolangCI <42910462+golangcibot@users.noreply.github.com>
Co-authored-by: Anca Zamfir <ancazamfir@users.noreply.github.com>
Co-authored-by: Marko <marbar3778@yahoo.com>
Co-authored-by: Anton Kaliaev <anton.kalyaev@gmail.com>
* Add processor prototype
* Change processor API
+ expose a simple `handle` function which mutates internal state
* processor tests
* fix gofmt and ohter golangci issues
* scopelint var on range scope
* add check for short block received
* fix formatting
* small test reorg
* ignore unused for now
* ci fix changes
* go.mod revert
+ use `trySend` the replicate peer sending
+ expose `next()` as a chan of events as output
+ expose `final()` as a chan of error, for the final error
+ add `ready()` as chan struct when routine is ready
+ ensure that we stop accepting messages once `stop` has been called
to avoid the case in which we attempt to write to a channel which
has already been closed
+ `routine.send` returns false when routine is not running
+ this will prevent panics sending to channels which have been
closed
+ Make output channels routine specific removing the risk of someone
writting to a channel which was closed by another touine.
+ consistency changes between the routines and the demuxer
Anton Kaliaev
Marko
Erik Grinaker
Sean Braithwaite