* Create an empty storage/db module
* Use ephemeral storage in tests
* Populate storage inside new() method
* Move scanner setup into an init() method
* Pass the network to scanner init
* Create a database but don't actually use it
* Skip shutdown format checks when skipping format upgrades
* Allow the scanner to skip launching format upgrades in production
* Refactor skipping format upgrades so it is consistent
* Allow checking configs for equality
* Restore Network import
* Provide access to DiskDb and DiskWriteBatch outside the state using a feature
* Actually let's export ZebraDb for the format upgrade code
* Pass column families to ZebraDb as an argument
* Allow the database kind to be changed in config.rs
* Use the state kind in finalized_state.rs
* Allow different database kinds in ZebraDb, but don't move the upgrade code yet
* Allow different database kinds in DiskDb
* Allow different database kinds in upgrade.rs, but don't split the upgrade code out yet
* Add new arguments to raw database tests
* Fix doc links
* Fix internal imports
* Fix unused code
* Update zebrad version metadata
* Create a specific state database delete function
* Fix state exports
* Fix zebrad tests
* Fix zebrad state write tests
* Make CI run again
* Fix dead code warnings for test methods
* Remove unnecessary async on some tests
* Fix logging required by tests
* Fix logging required in test itself
* Fix variable names
* Try to copy the message and add regexes
* Provide a user hint when the database format is invalid
* Split a path-based database version method
* Check the database format before Zebra shuts down
* Fix a typo in zebra-state/src/service/finalized_state/zebra_db.rs
Co-authored-by: Marek <mail@marek.onl>
* Fix a typo in upgrade.rs
---------
Co-authored-by: Marek <mail@marek.onl>
* If there's an existing database with no version file, give it version 25.0.0
* Creating the RocksDB database makes a temporary change to the default database version
* Move format upgrades to their own module and enum
* Launch a format change thread if needed, and shut it down during shutdown
* Add some TODOs and remove a redundant timer
* Regularly check for panics in the state upgrade task
* Only run example upgrade once, change version field names
* Increment database format to 25.0.2: add format change task
* Log the running and initial disk database format versions on startup
* Add initial disk and running state versions to cached state images in CI
* Fix missing imports
* Fix typo in logs workflow command
* Add a force_save_to_disk argument to the CI workflow
* Move use_internet_connection into zebrad_config()
* fastmod can_spawn_zebrad_for_rpc can_spawn_zebrad_for_test_type zebra*
* Add a spawn_zebrad_without_rpc() function
* Remove unused copy_state() test code
* Assert that upgrades and downgrades happen with the correct versions
* Add a kill_and_return_output() method for tests
* Add a test for new_state_format() versions (no upgrades or downgrades)
* Add use_internet_connection to can_spawn_zebrad_for_test_type()
* Fix workflow parameter passing
* Check that reopening a new database doesn't upgrade (or downgrade) the format
* Allow ephemeral to be set to false even if we don't have a cached state
* Add a test type that will accept any kind of state
* When re-using a directory, configure the state test config with that path
* Actually mark newly created databases with their format versions
* Wait for the state to be opened before testing the format
* Run state format tests on mainnet and testnet configs (no network access)
* run multiple reopens in tests
* Test upgrades run correctly
* Test that version downgrades work as expected (best effort)
* Add a TODO for testing partial updates
* Fix missing test arguments
* clippy if chain
* Fix typo
* another typo
* Pass a database instance to the format upgrade task
* Fix a timing issue in the tests
* Fix version matching in CI
* Use correct env var reference
* Use correct github env file
* Wait for the database to be written before killing Zebra
* Use correct workflow syntax
* Version changes aren't always upgrades
---------
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
* Implement minor and patch database format versions
* Log and update database format versions when opening database
* Refactor the current list of column families into a constant
* Open all available column families, including from future Zebra versions
* Refactor note commitment tree lookups to go through the height methods
* Make Sapling/Orchard note commitment tree lookup forwards compatible
* Ignore errors reading column family lists from disk
* Update format version comments and TODOs
* Correctly log newly created database formats
---------
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
* Rewrite some state cache docs to clarify
* Add a zebra_network::Config.cache_dir for peer address caches
* Add new config test files and fix config test failure message
* Create some zebra-chain and zebra-network convenience functions
* Add methods for reading and writing the peer address cache
* Add cached disk peers to the initial peers list
* Add metrics and logging for loading and storing the peer cache
* Replace log of useless redacted peer IP addresses
* Limit the peer cache minimum and maximum size, don't write empty caches
* Add a cacheable_peers() method to the address book
* Add a peer disk cache updater task to the peer set tasks
* Document that the peer cache is shared by multiple instances unless configured otherwise
* Disable peer cache read/write in disconnected tests
* Make initial peer cache updater sleep shorter for tests
* Add unit tests for reading and writing the peer cache
* Update the task list in the start command docs
* Modify the existing persistent acceptance test to check for peer caches
* Update the peer cache directory when writing test configs
* Add a CacheDir type so the default config can be enabled, but tests can disable it
* Update tests to use the CacheDir config type
* Rename some CacheDir internals
* Add config file test cases for each kind of CacheDir config
* Panic if the config contains invalid socket addresses, rather than continuing
* Add a network directory to state cache directory contents tests
* Add new network.cache_dir config to the config parsing tests
* ZIPs were updated to remove ambiguity, this was tracked in #1267.
* #2105 was fixed by #3039 and #2379 was closed by #3069
* #2230 was a duplicate of #2231 which was closed by #2511
* #3235 was obsoleted by #2156 which was fixed by #3505
* #1850 was fixed by #2944, #1851 was fixed by #2961 and #2902 was fixed by #2969
* We migrated to Rust 2021 edition in Jan 2022 with #3332
* #1631 was closed as not needed
* #338 was fixed by #3040 and #1162 was fixed by #3067
* #2079 was fixed by #2445
* #4794 was fixed by #6122
* #1678 stopped being an issue
* #3151 was fixed by #3934
* #3204 was closed as not needed
* #1213 was fixed by #4586
* #1774 was closed as not needed
* #4633 was closed as not needed
* Clarify behaviour of difficulty spacing
Co-authored-by: teor <teor@riseup.net>
* Update comment to reflect implemented behaviour
Co-authored-by: teor <teor@riseup.net>
* Update comment to reflect implemented behaviour when retrying block downloads
Co-authored-by: teor <teor@riseup.net>
* Update `TODO` to remove closed issue and clarify when we might want to fix
Co-authored-by: teor <teor@riseup.net>
* Update `TODO` to remove closed issue and clarify what we might want to change in future
Co-authored-by: teor <teor@riseup.net>
* Clarify benefits of how we do block verification
Co-authored-by: teor <teor@riseup.net>
* Fix rustfmt errors
---------
Co-authored-by: teor <teor@riseup.net>
* Implement Display and to_string() for NetworkUpgrade
* Add a progress-bar feature to zebrad
* Add the progress bar writer to the tracing component
* Add a block progress bar transmitter
* Correctly shut down the progress bar, and shut it down on an interrupt
* Make it clearer that the progress task never exits
* Add a config for writing logs to a file
* Add a progress-bar feature to zebra-network
* Add a progress bar for the address book size
* Add progress bars for never attempted and failed peers
* Add an optional limit and label to connection counters
* Add open connection progress bars
* Improve CheckpointList API and CheckpointVerifier debugging
* Add checkpoint index and checkpoint queue progress bars
* Security: Limit the number of non-finalized chains tracked by Zebra
* Make some NonFinalizedState methods available with proptest-impl
* Add a non-finalized chain count progress bar
* Track the last fork height for newly forked chains
* Add a should_count_metrics to Chain
* Add a display method for PartialCumulativeWork
* Add a progress bar for each chain fork
* Add a NonFinalizedState::disable_metrics() method and switch to using it
* Move metrics out of Chain because we can't update Arc<Chain>
* Fix: consistently use best chain order when searching chains
* Track Chain progress bars in NonFinalizedState
* Display work as bits, not a multiple of the target difficulty
* Handle negative fork lengths by reporting "No fork"
* Correctly disable unused fork bars
* clippy: rewrite using `match _.cmp(_) { ... }`
* Initial mempool progress bar implementation
* Update Cargo.lock
* Add the actual transaction size as a description to the cost bar
* Only show mempool progress bars after first activation
* Add queued and rejected mempool progress bars
* Clarify cost note is actual size
* Add tracing.log_file config and progress-bar feature to zebrad docs
* Derive Clone for Chain
* Upgrade to howudoin 0.1.2 and remove some bug workarounds
* Directly call the debug formatter to Display a Network
Co-authored-by: Arya <aryasolhi@gmail.com>
* Rename the address count metric to num_addresses
Co-authored-by: Arya <aryasolhi@gmail.com>
* Simplify reverse checkpoint lookup
Co-authored-by: Arya <aryasolhi@gmail.com>
* Simplify progress bar shutdown code
Co-authored-by: Arya <aryasolhi@gmail.com>
* Remove unused MIN_TRANSPARENT_TX_MEMPOOL_SIZE
* Document that the progress task runs forever
* Fix progress log formatting
* If progress-bar is on, log to a file by default
* Create missing directories for log files
* Add file security docs for running Zebra with elevated permissions
* Document automatic log file, spell progress-bar correctly
---------
Co-authored-by: Arya <aryasolhi@gmail.com>
* add initial ES support
* hide elasticsearch behind feture, add config
* make the builds and clippy happy
* move ES code to a function
* change database name
* fix database name and panic
* increase close to tip interval
* update deny for elastic
* remove a block clone
* try to fix builds
* fix builds 2
* refactor some imports and unwraps
* typo
* change argument of elasticsearch function
* move elastic call to the end of `commit_finalized_direct`
* get height from block
* delete old database directories
* check if state directory exists
* skip deleting when ephemeral
* split `check_and_delete_old_databases`
* move `check_and_delete_old_databases` to state
* spawn `check_and_delete_old_databases`
* simplity a bit
* fix(state): only delete old database directories inside the cache directory (#4631)
* Add function comments, tweak log
* Simplify version parsing
* Use spawn_blocking to launch the task on a separate thread, do the cleanup last
* Abort the cleanup task when Zebra exits
* Split directory deletion into its own function, handle ownership
* Rename cache_dir to state_dir
* If an outdated state directory is outside the cache directory, don't delete it
* Minimise diffs
* add test
* fix typos
Co-authored-by: teor <teor@riseup.net>
* add `canonicalize` to test regex
* add another match to test
Co-authored-by: teor <teor@riseup.net>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
* Fix state directory and docs
* Explain state versions better, and fix a docs bug
* fix docs
Co-authored-by: Marek <mail@marek.onl>
* add extra info to doc
Co-authored-by: Alfredo Garcia <oxarbitrage@gmail.com>
Co-authored-by: Marek <mail@marek.onl>
* Export the `zebra_state::Config::db_path` method
Make it easier for tests to discover the sub-directory used to store
the chain state data.
* Generate code for interfacing with lightwalletd
Use the `tonic-build` crate to generate Rust code for communicating with
lightwalletd using gRPC.
The `*.proto` files were obtained from the Zcash lightwalletd
repository.
* Use `block::Height` instead of `Height`
Import the `block` instead to make it slightly clearer.
* Add helper function to remove a file if it exists
Try to remove it and ignore an error if it says that the file doesn't
exist. This will be used later to remove the lock file from a copied
chain state directory.
* Add helper function to copy chain state dirs
Copy an existing chain state directory into a new temporary directory.
* Add a `BoxStateService` type alias
Make it easier to write and read a boxed version of a state service.
* Add a helper function to start the state service
Make it easier to specify the state service to use an existing state
cache directory.
* Import `eyre!` macro at the module level
Allow it to be used in different places without having to repeat the
imports.
* Add `load_tip_height_from_state_directory` helper
A function to discover the current chain tip height stored in a state
cache.
* Add helper function to prepare partial sync. state
Loads a partially synchronized cached state directory into a temporary
directory that can be used by a zebrad instance, and also returns the
chain tip block height of that state.
* Add `perform_full_sync_starting_from` helper
Runs a zebrad with an existing partially synchronized state, and
finishes synchronizing it to the network chain tip.
* Add function to load transactions from a block
Use a provided state service to load all transactions from a block at a
specified height.
The state service is a generic type parameter, because
`zebra_state::service::ReadStateService` is not exported publicly. Using
a generic type parameter also allows the service to be wrapped in layers
if needed in the future.
* Add `load_transactions_from_block_after` helper
A function to load transactions from a block stored in a cached state
directory. The cached state must be synchronized to a chain tip higher
than the requested height.
* Add helper function to load some test transactions
Given a partially synchronized chain state, it will extend that chain by
performing a full synchronization, and obtain some transactions from one
of the newly added blocks.
* Update `spawn_zebrad_for_rpc_without_initial_peers`
Wait until the mempool is activated.
* Add method to start lightwalletd with RPC server
Returns the lightwalletd instance and the port that it's listening for
RPC connections.
The instance can reuse an existing cached lightwalletd state if the
`LIGHTWALLETD_DATA_DIR` environment variable is set.
* Add a `LightwalletdRpcClient` type alias
To make it easier to identify the type generated from the Protobuf
files.
* Add helper function to connect to lightwalletd
Prepare an RPC client to send requests to a lightwalletd instance.
* Add a `prepare_send_transaction_request` helper
Creates a request message for lightwalletd to send a transaction.
* Add test to send transactions using lightwalletd
Obtain some valid transactions from future blocks and try to send them
to a lightwalletd instance connected to a zebrad instance that hasn't
seen those transactions yet. The transactions should be successfully
queued in Zebra's mempool.
* Make `zebra_directory` parameter generic
Allow using a `TempDir` or a `PathBuf`.
* Move lightwalletd protobuf files
Place them closer to the module directory, so that it's clearer that
they specify the RPC protocol for lightwalletd, and not Zebra itself.
* Don't use coinbase transactions in the test
Coinbase transactions are rejected by the mempool.
* Don't remove state lock file
It is removed automatically by Zebra when it shuts down, so if it exists
it should be reported as a bug.
* Force mempool to be enabled in Zebrad instance
Speed up the initialization of the Zebrad instance used for lightwalletd
to connect to.
* Refactor to create `LIGHTWALLETD_DATA_DIR_VAR`
Document how the environment variable can be used to speed up the test.
* Check for process errors in spawned Zebra instance
Enable checking for known process failure messages.
* Add `FINISH_PARTIAL_SYNC_TIMEOUT` constant
Document why it exists and how the choice of the value affects the test.
* Add `LIGHTWALLETD_TEST_TIMEOUT` constant
And use it for the Zebrad and the Lightwalletd instances used in the
send transaction integration test.
* Check `lightwalletd` process for errors
Enable checking the lightwalletd process for known failure messages.
* Update `tonic` and `prost` dependencies
Use the latest version and fix CI failures because `rustfmt` isn't
installed in the build environment.
* Create `send_transaction_test` module
Move the send transaction using lightwalletd test and its helper
functions into a new module.
* Move `LIGHTWALLETD_TEST_TIMEOUT` constant
Place it in the parent `lightwalletd` module.
* Move gRPC helper functions and types to `rpc` mod.
Make them more accessible so that they can be used by other tests.
* Create a `cached_state` module
Move the test utility functions related to using a cached Zebra state
into the module.
* Move `perform_full_sync_starting_from` to `sync`
Keep to closer to the synchronization utility functions.
* Move Zebra cached state path variable constant
Place it in the `cached_state` module.
* Skip test if `ZEBRA_TEST_LIGHTWALLETD` is not set
Make it part of the set of tests ignored as a whole if no lightwalletd
tests should be executed.
* Move `spawn_zebrad_for_rpc_without_initial_peers`
Place it in the `launch` sub-module.
* Rename `rpc` module into `wallet_grpc`
Avoid any potential misunderstandings when the name is seen out of
context.
* Allow duplicate `heck` dependency
At least until `structopt` is updated or `zebra-utils` is updated to use
`clap` 3.
* Fix a deny.toml typo
* fix(build): CMake is required by `prost` crate
Co-authored-by: teor <teor@riseup.net>
Co-authored-by: Gustavo Valverde <gustavo@iterativo.do>
* refactor(state): move RocksDB-specific initialization to a new module
* refactor(state): move RocksDB-specific shutdown to a new module
* refactor(state): temporarily allow RocksDB-specific reads and writes, without a new module
Unlike the last few commits, this one actually compiles.
* refactor(state): add a DiskWriteBatch wrapper for RocksDB writes
* refactor(state): move finalized state test methods to a test module
Also:
- upgrades to rlimit 0.7.0
- updates types to match the breaking changes in rlimit
- deletes a manual implementation that was similar to `increase_nofile_limit`,
but not as good on macOS and some BSDs
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
* Use `testdir()` instead of `TempDir::new()`
Reduce repeated code and make it easier to change from using `tempdir`
to use `tempfile` instead.
* Replace `tempdir` with `tempfile` in `zebrad`
Use `tempfile`'s `TempDir` instead.
* Use `tempdir()` instead of `TempDir::new()`
Reduce repeated code and make it easier to upgrade to `tempfile`.
* Use `tempfile` instead of `tempdir`
Replace obsoleted `tempdir` dependency with `tempfile`.
* Use `tempfile` instead of `tempdir`
Replace obsoleted `tempdir` dependency with `tempfile`.
* Update `Cargo.lock`
Update it now that `tempdir` has been replaced with `tempfile`.
* Remove `tempdir` from `deny.toml` exceptions
Ban duplicate versions of the `tempdir` dependency.
* Remove `inferno` from `deny.toml` exceptions
It apparently isn't needed anymore.
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Deirdre Connolly <deirdre@zfnd.org>
If the limit is less than the ideal, try to increase it to the ideal.
If that doesn't work, try to increase the limit as high as possible.
If the limit is still less than the minimum, panic.
Some systems have a very small /dev/shm, for example, see:
https://github.com/docker-library/postgres/issues/416
So we should just use the temporary directory on all operating systems.
Also:
* use TempDir to generate the temporary path
* delete the code that we copied from sled
* prefix the temporary path with the state version and network
## Motivation
Prior to this PR we've been using `sled` as our database for storing persistent chain data on the disk between boots. We picked sled over rocksdb to minimize our c++ dependencies despite it being a less mature codebase. The theory was if it worked well enough we'd prefer to have a pure rust codebase, but if we ever ran into problems we knew we could easily swap it out with rocksdb.
Well, we ran into problems. Sled's memory usage was particularly high, and it seemed to be leaking memory. On top of all that, the performance for writes was pretty poor, causing us to become bottle-necked on sled instead of the network.
## Solution
This PR replaces `sled` with `rocksdb`. We've seen a 10x improvement in memory usage out of the box, no more leaking, and much better write performance. With this change writing chain data to disk is no longer a limiting factor in how quickly we can sync the chain.
The code in this pull request has:
- [x] Documentation Comments
- [x] Unit Tests and Property Tests
## Review
@hdevalence
Closes#1026
Because of the way that sled uses this parameter, the actual in-memory
size may be much larger. Dialing this down should help avoid high
memory usage.
* Run large checkpoint sync tests in CI
* Improve test child output match error context
* Add a debug_stop_at_height config
* Use stop at height in acceptance tests
And add some restart acceptance tests, to make sure the stop at
height feature works correctly.
teor
Pili Guerra
Alfredo Garcia
Janito Vaqueiro Ferreira Filho
Jane Lusby
Henry de Valence
Jane Lusby