Skip to content

Latest commit

 

History

History
267 lines (218 loc) · 19.2 KB

CHANGELOG.md

File metadata and controls

267 lines (218 loc) · 19.2 KB

Change log

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning. See MAINTAINERS.md for instructions to keep up to date.

Unreleased

System Administration Changes

Added

  • Added support for EOSIO 2.1.x (requires 2.1.1 to have full data since some data where bogus in 2.1.0)
  • Added support for environment variable EOSWS_PUSH_V1_OUTPUT=true that forces push-transaction (guarantee:in-block) to output the same content format as nodeos 2.0.x (with Inlines)
  • Added support for environment variable DSTORE_S3_BUFFERED_READ=true that forces reading S3 files (ex: blocks) ahead of processing, useful when S3 provider has trouble keeping long connections alive.
  • Added support for looking up irreversible blocks on blockmeta (when the LIB was stuck for a while) from tokenmeta and trxdb-loader. They now use '--common-blockmeta-addr' flag if available
  • Added --common-chain-core-symbol flag to define actual chain core symbol in the form <precision>,<symbol code> defaults to 4,EOS by default.
  • Added --tokenmeta-readiness-max-latency with default=5m, now tokenmeta will show as "NotServing" through grpc healthcheck if last processed block (HEAD) is older than this. Value of 0 disables that feature.
  • Added --relayer-source-request-burst with default=90 to allow a relayer connecting to another relayer to request a 'burst'
  • Added --statedb-disable-indexing to disable indexing of tablet and injecting data into storage engine developer option, don't use that in production.
  • Added --eosws-nodeos-rpc-push-extra-addresses to allow providing a list of backup EOS addresses when push-transaction does not succeed in getting the transaction inside a block (with push_guarantee)
  • Added --eosws-max-stream-per-connection to allow changing how many stream can be active at the same time for a given WebSocket connection, defaults to 12 which was the hard-coded value.
  • Added --eosws-statedb-proxy-retries, Number of time to retry proxying a request to statedb before failing (default 2)
  • Added --eosws-nodeos-rpc-proxy-retries, Number of time to retry proxying a request to statedb before failing (default 2)
  • Added --mindreader-max-console-length-in-bytes which is the limit in bytes that we allow action trace's console output to be before truncating them.
  • Environment variable MINDREADER_MAX_TOKEN_SIZE can now be set to override bufio.Scanner() max token size (default 52428800, i.e. 50Mb) for EOSIO chains with huge transactions
  • Flag --accounthist-mode to specific the accounthist mode of operation
  • Added tools check accounthist-shards to
  • Flag --common-include-filter-expr, --common-exclude-filter-expr, --common-system-actions-include-filter-expr can optionally specify multiple values, separated by ;;; and prefixed by #123; where 123 is a block number at which we stat applying that filter
  • Added accounthist tools allows you to scan and read accounts dfuseeos tools accounthist read ... dfuseeos tools accounthist scan ...
  • Flag --search-router-truncation-low-block-num to make the router aware of lower-block-truncation and serve requests accordingly
  • Flag --mindreader-oneblock-suffix that mindreaders can each write their own file per block without competing for writes. #140
  • Flag --eosws-disabled-messages a comma separated list of ws messages to disable.
  • Flag --common-system-shutdown-signal-delay, a delay that will be applied between receiving SIGTERM signal and shutting down the apps. Health-check for eosws and dgraphql will respond 'not healthy' during that period.

Removed

  • Breaking Change Removed --eosq-price-ticker-name flag, if you were using this flag, please use --common-chain-core-symbol instead to define it.
  • Removed dgraphql-graceful-shutdown-delay, it was a left-over, unused. Must use --common-system-shutdown-signal-delay now
  • Removed relayer-max-drift (now dependent on a new condition of presence of a "block hole", and no new block sent for 5 seconds)
  • Removed relayer-init-time (no need for it with this new condition ^)

Changed

  • The --eosq-available-networks logo field each network now has a maximum height of 70px.
  • The --eosq-available-networks config of each network now accepts a logo_text that when present, is displayed alongside the logo field. This field is taken into consideration only when logo is defined. In this mode, the logo is fixed to 48px x 48px. If the logo_text value is eosq, this is rendered like the standard eosq logo.
  • Breaking Change Changes to --eosq-available-networks config might be required around the logo field each network. You must now remove the logo field if it's not pointing to an existing image otherwise, the logo will not be rendered correctly.
  • Applying a block filter over previously-filtered-blocks does not panic anymore, it applies the new filter on top of it, only if that specific filter has never been applied before. Applied filters definitions are concatenated in the block metadata, separated by ;;;
  • Default trxdb-loader-batch-size changed to 100, Safe to do so because it does not batch when close to head.
  • Improved relayer mechanics: replaced "max drift" detection by "block hole" detection and recovery action is now to restart the joining source (instead of shutting down the process)
  • Improved dfuseeos tools check statedb-reproc-injector output by showing all shard statistics (and not just most highest block).
  • Breaking Change Changed --statedb-enable-pipeline flag to --statedb-disable-pipeline to make it clearer that it should not be disable, if you were using the flag, change the name and invert the logical value (i.e. --state-enable-pipeline=false becomes --state-disable-pipeline=true)

Fixed

  • Fixed Firehose gRPC listening address over plain text.
  • Fixed validation of transaction ID passed to WebSocket get_transaction API, the prior validation was too permissive.
  • Fixed a bug making search-forkresolver useless, because ignored by search-router.
  • Fixed a bug on StateDB server not accepting symbol and symbol code as scope parameter value.
  • Fixed shutdown on dgraphql (grpc/http) so it closes the active connections a little bit more gracefully.
  • Fixed a bug in TiKV store implementation preventing it to delete keys correctly.
  • Fixed a bug in eosws WebSocket get_transaction_lifecycle where a transaction not yet in the database would never stream back any message to the client.
  • Fixed a bug with --mindreader-no-blocks-log option actually not being picked up (always false)
  • Fixed a bug with /state/table/row not correctly reading row when it was in the table index.
  • Fixed a bug with /state/tables/scopes where the actual block num used to query the data was incorrect leading to invalid response results.
  • Fixed a bug with gRPC dfuse.eosio.statedb.v1/State#StreamMultiScopesTableRows where the actual block num used to query the data was incorrect leading to invalid response results.
  • Fixed issue when reading ABI from StateDB where speculative writes were not handled correctly.
  • Fixed issue when reading Table Row from StateDB where speculative writes were not handled correctly.
  • Fixed a potential crash when reading ABI from StateDB and it does not exist in database.

[v0.1.0-beta8] 2020-08-08

  • fix experimental netkv implementation for statedb

[v0.1.0-beta7] 2020-12-07

  • fix experimental netkv implementation for trxdb

[v0.1.0-beta6] 2020-08-27

  • When using filtering capabilities, only absolutely required system actions will be indexed/processed.
  • Added missing updateauth and deleteauth as require system actions in flag common-system-actions-include-filter-expr.

[v0.1.0-beta5] 2020-08-24

PUBLIC API Changes

Added

  • Added tokens, accountBalances, tokenBalances calls to dgraphql (based on tokenmeta)

System Administration Changes

Changed

  • Breaking Change FluxDB has been extracted to a dedicated library (github.com/streamingfast/fluxdb) with complete re-architecture design.
  • Breaking Change FluxDB has been renamed to StateDB and is incompatible with previous written data. See FluxDB Migration section below for more details on how to migrate.
  • merger startblock behavior changed, now relies on state-file, see streamingfast/merger#1
  • Changed merger-seen-blocks-file flag to merger-state-file to reflect this change.
  • merger now properly handles storage backend errors when looking for where to start
  • mindreader now automatically produces "merged blocks" instead of "one-block-files" when catching up (based on blocktime or if a blockmeta is reachable at --common-blockmeta-addr)
  • mindreader now sets optimal EOS VM settings automatically if the platform supports it them when doing dfuseeos init.
  • Changed --abicodec-export-cache-url flag to abicodec-export-abis-base-url and will contain only the URL of the where to export the ABIs in JSON.
  • Changed --abicodec-export-cache flag to abicodec-export-abis-enabled.

Added

  • Added tokenmeta app, with its flags
  • Added support for filtered blocks, search indices and trxdb, with --common-include-filter-expr and --common-include-filter-expr.
  • Added merged-filter app (not running by default), that generates filtered merged blocks files from regular merged blocks files.
  • Added truncation handling to trxdb-loader, which will only keep a moving window of data in trxdb, and delete preceding transactions. Enabling that feature requires reprocessing trxdb.
    • --trxdb-loader-truncation-enabled
    • --trxdb-loader-truncation-window
    • --trxdb-loader-truncation-purge-interval
  • Added --statedb-reproc-shard-scratch-directory to run StateDB reprocessing sharder using scratch directory to reduce RAM usage
  • Added --merger-one-block-deletion-threads (default:10) to allow control over one-block-files deletion parallelism
  • Added --merger-max-one-block-operations-batch-size to allow control over one-block-files batches that are looked up on storage,
  • Added --eosws-with-completion (default: true) to allow control over that feature
  • Added --mindreader-merge-threshold-block-age when processing blocks with a blocktime older than this threshold, they will be automatically merged: (default 12h)
  • Added --mindreader-batch-mode to force always merging blocks (like --mindreader-merge-and-store-directly did) AND overwriting existing files in destination.
  • Added --mindreader-wait-upload-complete-on-shutdown flag to control how mindreader waits on upload completion when shutting down (previously waited indefinitely)
  • Added --search-live-hub-channel-size flag to specific the size of the search live hub channel capacity
  • Added --search-live-preprocessor-concurrent-threads: number of thread used to run file source preprocessor function
  • Added --abicodec-export-abis-file-name, contains the URL where to export the ABIs in JSON
  • Added --metrics-listen-addr to control on which address to server the metrics API (Prometheus), setting this value to an empty string disable metrics serving.
  • Added --dashboard-metrics-api-addr to specify a different API address where to retrieve metrics for the dashboard.
  • Added Experimental support for kvdb backend netkv://, an extremely simple network layer over badger to allow running dfuse components in separate instances.

Removed

  • The --merger-delete-blocks-before flag is now removed and is the only behavior for merger.
  • The --mindreader-merge-and-store-directly flag was removed. That behavior is now activated by default when encountering 'old blocks'. Also see new flag mindreader-batch-mode.
  • The --mindreader-discard-after-stop-num flag was removed, its implementation was too complex and it had no case where it was really useful.
  • The --mindreader-producer-hostname flag was removed, this option made no sense in the context of mindreader app.
  • The --eosq-disable-tokenmeta flag was removed, token meta is now included, so this flag is now obsolete.
  • The --eosq-on-demand flag was removed, this was unused in the codebase.

Fixed

  • Fixed issue where blockmeta was not serving on GRPC at all because it couldn't figure out where to start on the stream
  • Fixed issue with merger with a possible panic when reading a one-block-file that is empty, for example on a non-atomic storage backend
  • Fixed issue with mindreader not stopping correctly (and showing any error) if the bootstrap phase (ex: restore-from-snapshot) failed.
  • Fixed issue with pitreos not taking a backup at all when sparse-file extents checks failed.
  • Fixed issue with dfuseeos tools check merged-blocks (start/end block, false valid ranges when the first segment is not 0, etc.)
  • Improved performance by using value for bstream.BlockRef instead of pointers and ensuring we use the cached version.
  • mindreader and node-manager improved nodeos log handling

FluxDB to StateDB Migration

FluxDB required an architecture re-design to fit with our vision about the tool and make it chain agnostic (so it is easier to re-use on our other supported chain).

The code that was previously found here has been extracted to its own library (https://github.com/streamingfast/fluxdb). There is now a new app named StateDB (statedb is the app identifier) in dfuse for EOSIO that uses FluxDB to support all previous API endpoints served by the FluxDB app as well as now offering a gRPC interface.

While doing this, we had to change how keys and data were written to the underlying engine. This means that all your previous data stored cannot be read anymore by the new StateDB and that new data written by StateDB will not be compatible on a previous instance.

What that means exactly is that StateDB will require to re-index all the current merged blocks up to live blocks before being able to serve requests. This is the main reason why we decided to rename the app, so you are forced to peform this step.

Here the steps required to migrate to the new statedb app:

  1. We strongly recommend that you take a full backup of your data directory (while the app is shut down)
  2. Launch a stand-alone stateDB instance in 'inject-mode' that reads from your block files and writes to a new location (see '--statedb-store-dsn')
  3. Let it complete the "catch up" until it is very close to the HEAD of your network, then stop that instance.
  4. Stop your previous instance (that uses fluxdb),
  5. Copy the content of your statedb database to a location accessible from there (that you will define in '--statedb-store-dsn')
  6. Launch the new version of the code, with the modified flags, over your previous data, including the new statedb database content (see below for the necessary flag and config modifications)

Here are the flags and config modifications required for switching to fluxdb to statedb, once the new data has been generated.

  • In your dfuse.yaml config, under replace the fluxdb app by statedb and all flags prefixed with fluxdb- must now be prefixed with statedb-.

    From:

    start:
    args:
    - ...
    - fluxdb
    - ...
    flags:
      ...
      fluxdb-http-listen-addr: :9090
      ...

    To:

    start:
    args:
    - ...
    - statedb
    - ...
    flags:
      ...
      statedb-http-listen-addr: :9090
      ...
  • If you had a customization for fluxdb-statedb-dsn, you must first renamed it statedb-store-dsn to statedb-store-dsn. Important You must use a new fresh database, so update your argument so it points to a new database, ensuring we don't overwrite data over old now incompatible data. If you did not customize the flag, continue reading, the default value has changed to point to a fresh storage folder.

    If you had a customization for eosws-flux-addr, rename to eosws-statedb-grpc-addr and ensure it points to the StateDB GRPC address (and not the HTTP address), its value must be the same as the flag statedb-grpc-listen-addr.

  • If you have a custom fluxdb-max-threads flag, removed it, customizing this value is not supported anymore.

[v0.1.0-beta4] 2020-06-23

See release notes.

[v0.1.0-beta3] 2020-05-13

See release notes.

[v0.1.0-beta2] 2020-04-27

Added

  • Added apiproxy application, with its flags
  • Added --log-format option for JSON output and log-to-file bool (default to true, same behavior as before)
  • Filtering (whitelist and blacklist) of what is indexed in Search, based on Google's Common Expression Language. See details here. Added --search-common-action-filter-on-expr and --search-common-action-filter-out-expr.
    • NOTE: This doesn't affect what is extracted from the chain, allowing you to re-index selectively without a chain replay.

Changed

  • CLI: dfuseeos init now writes dfuse.yaml with the start command's flags, also the array of components to start
  • CLI: new {dfuse-data-dir} replacement string in config flags, also changed default flag values
  • --node-manager-config-dir now ./producer (was manager/config)
  • --node-manager-data-dir now {dfuse-data-dir}/node-manager/data (was managernode/data)
  • --mindreader-config-dir now ./mindreader (was mindreadernode/config)
  • --mindreader-data-dir now {dfuse-data-dir}/mindreader/data (was mindreadernode/data)
  • CLI: regrouped some flags:
  • --search-indexer-dfuse-hooks-action-name, --search-live-dfuse-hooks-action-name, --search-forkresolver-dfuse-hooks-action-name fused into new --search-common-dfuse-hooks-action-name.
  • --search-...-mesh-publish-polling-duration fused into new --search-common-mesh-publish-polling-duration.
  • all of the --search-mesh-... options were renamed to --search-common-mesh-... (previously --search-mesh-service-version, --search-mesh-namespace, --search-mesh-store-addr)
  • dashboard: now separate metrics for mindreader vs producer node
  • dashboard doesn't act as a reverse proxy anymore (apiproxy does)
  • dashboard's default port is now :8081
  • eosq's port is now proxied through :8080, so use that.
  • App manager renamed to node-manager. All of its flags were changed from --manager-... to --node-manager-...

Removed

  • The --search-...-indexing-restrictions-json. This was replaced by the filtering listed above.

[v0.1.0-beta1] 2020-04-17

Added

  • Added --nodeos-path to control which nodeos executable is used.

Removed

  • Removed --chain-name, replaced by --config-file (or -c)
  • Removed init --reset option, dfusebox purge does it all now.

Changed

  • Renamed dfusebox to dfuse for EOSIO
    • dfusebox.yaml to dfuse.yaml
    • dfusebox-data to dfuse-data
  • --data-dir now defaults to ./dfusebox-data, and is separate from a chain name or the config file location.
  • dfusebox init now only generates a dfusebox.yaml config file, which can be booted with dfusebox start.
  • dfusebox init now only has the interactive method. We can later add more programmatic method to boot chains. With a dfusebox.yaml config now, however, we can reuse initializations multiple times.
  • License changed to Apache 2.0
  • Added GitHub workflow for PR checks