172 Commits (ea3da7fd079acb28f7cf54a93b48cf1ee5324933)

Author	SHA1	Message	Date
Luke Imhoff	b1f4696569	Reconnect, rerequest, and resubscribe when websocket disconnects ... * Use struct, `%EthereumJSONRPC.WebSocket.WebSocketClient.Options{}` to hold `web_socket_options`. * `%EthereumJSONRPC.WebSocket.Registration{}` * Record `t:EthereumJSONRPC.request/0` in `%EthereumJSONRPC.WebSocket.Registration{}`, so that the request can be re-requested if the websocket disconnects and reconnects. * Drop server-side `subscription_id` from `%EthereumJSONRPC.WebSocket.Registration{}` as it can retrieved from `request` now. * `EthereumJSONRPC.WebSocket.WebSocketClient` * Track whether the web socket is connected or not, so that requests aren't sent when in the disconnected state. This prevents double-sends as `handle_call` can be called before `onconnect`, which caused that first subscribe to be sent twice. * Instead of the subscriber knowing the server-side subscription id, an opaque `t:reference/0` is used, so that the reference is stable in the subscriber's `%Subscription{}`, but the subscription ID can change on reconnect, which means replacing `subscription_id_to_subscription` in `%WebSocketClient{}` with 3 maps 1. `subscription_id_to_subscription_reference` - used translate server-side, per-connection subscription to the stable reference in `subscription_reference_to_subscription` 2. `subscription_reference_to_subscription_id` - used to convert stable reference to subscription id to unsubscribe. 3. `subscription_reference_to_subscription` - used to get full Subscription, including the subscriber_pid for when subscription messages are delivered or the subscription needs to be resubscribed using the `web_socket_options`. * `ondisconnect`, `:reconnect` immediately. * When disconnecting, tell unsubscribers they were successful. * Hold any json_rpc or subscribe requests to retry on reconnect. * Hold any subscriptions to resubscribe on reconnect. * `onconnect` * re-request the saved json_rpc and subscribe requests. * issue new requests to re-establish the subscriptions that had already succeeded in their subscribe requests. * `%EthereumJSONRPC.Subscription{}` * Replace `id`, which did match the server created ID, with a `t:reference/0`, which allows for indirection and for the server created ID to change on reconnect while the subscriber does not need to care about the reconnections. * `EthereumJSONRPC.WebSocket` * `options` is switched from a Keyword.t to a struct, `t`, so that it is easier to pattern match for the reconnect logic in `WebSocketClient`. * `web_socket_options` is no longer a generic `term` because the code did require it to be a map with a `:web_socket` key for the `t:GenServer.server/0`.	6 years ago
Konstantin Zolotarev	0604f7a234	Fixes for pr comments	6 years ago
Konstantin Zolotarev	f89a27adf4	Fixes for ganache integration	6 years ago
Luke Imhoff	8bb151da71	Log and retry import errors in uncle fetcher	6 years ago
Luke Imhoff	fcf5e6ef99	InternalTransaction.Fetcher buffer tuples instead maps ... ```elixir test "memory" do InternalTransaction.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(InternalTransaction.Fetcher) Stream.repeatedly(fn -> %{block_number: block_number(), hash: transaction_hash()} end) |> Enum.take(100) |> InternalTransaction.Fetcher.async_fetch() Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "AFTER GC") end ``` | Balances | Garbage Collection | Map (bytes) | Tuple (bytes) | Tuple / Map (%) | |----------|--------------------|-------------|---------------|-----------------| | 10,000 | Before | 4,119,668 | 2,546,460 | 61.81 | | 10,000 | After | 1,574,172 | 973,260 | 61.82 | | 100,000 | Before | 11,519,724 | 9,600,436 | 83.34 | | 100,000 | After | 11,516,716 | 9,597,428 | 83.33 |	6 years ago
Luke Imhoff	18e06d14f7	Indexer.CoinBalance.Fetcher tuple instead of map in buffers ... ```elixir test "memory" do CoinBalance.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(CoinBalance.Fetcher) Stream.repeatedly(fn -> %{address_hash: address_hash(), block_number: block_number()} end) |> Enum.take(100) |> CoinBalance.Fetcher.async_fetch_balances() Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "AFTER GC") end ``` | Balances | Garbage Collection | Map (bytes) | Tuple (bytes) | Tuple / Map (%) | |----------|--------------------|-------------|---------------|-----------------| | 1,000 | Before | 372,444 | 230,484 | 61.88 | | 1,000 | After | 142,924 | 88,636 | 62.02 | | 100,000 | Before | 11,519,724 | 9,600,436 | 83.34 | | 100,000 | After | 11,516,716 | 9,597,428 | 83.33 |	6 years ago
Luke Imhoff	f04ea541d0	Indexer.CoinBalance.Fetcher bytes and integer instead of string in ... buffers ```elixir test "memory" do CoinBalance.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(CoinBalance.Fetcher) Stream.repeatedly(fn -> %{address_hash: address_hash(), block_number: block_number()} end) |> Enum.take(100) |> CoinBalance.Fetcher.async_fetch_balances() Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> elem(1) |> IO.inspect(label: "AFTER GC") end ``` | Balances | Garbage Collection | String (bytes) | Bytes and Integer (bytes) | Bytes / String (%) | |----------|--------------------|----------------|---------------------------|--------------------| | 100 | Before | 55,244 | 34,548 | 62.54 | | 100 | After | 21,764 | 13,860 | 63.68 | | 1,000 | Before | 601,972 | 372,444 | 61.87 | | 1,000 | After | 230,596 | 142,924 | 61.98 | | 10,000 | Before | 4,119,668 | 2,546,460 | 61.81 | | 10,000 | After | 1,574,172 | 973,260 | 61.83 | | 100,000 | Before | 16,586,644 | 11,519,724 | 69.45 | | 100,000 | After | 16,583,636 | 11,516,716 | 69.45 |	6 years ago
Luke Imhoff	8d542f07bd	Buffer tuples instead of maps ... ```elixir test "memory" do TokenBalance.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(TokenBalance.Fetcher) 1_000 |> build_list(:token_balance) |> Enum.map(& %{&1 | address_hash: &1.address.hash, token_contract_address_hash: address_hash()}) |> TokenBalance.Fetcher.async_fetch() Process.info(pid, :memory) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> IO.inspect(label: "AFTER GC") end ``` | Token Balances | Garbage Collection | Map (bytes) | Tuple (bytes) | Tuple / Map (%) | |----------------|--------------------|-------------|---------------|-----------------| | 1,000 | Before | 233,500 | 145,828 | 62.45 | | 1,000 | After | 230,492 | 142,820 | 61.96 | | 50,000 | Before | 8,001,028 | 6,668,188 | 83.34 | | 50,000 | After | 7,998,020 | 6,665,180 | 83.35 |	6 years ago
Luke Imhoff	b313af6d7f	Buffer hash bytes instead of hash strings ... ```elixir test "memory" do TokenBalance.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(TokenBalance.Fetcher) 10_000 |> insert_list(:token_balance) |> TokenBalance.Fetcher.async_fetch() Process.info(pid, :memory) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> IO.inspect(label: "AFTER GC") end ``` | Token Balances | Garbage Collection | Hash Strings (bytes) | Hash Bytes (bytes) | Bytes / String (%) | |----------------|--------------------|----------------------|--------------------|--------------------| | 1,000 | Before | 233,500 | 233,500 | 100.00 | | 1,000 | After | 233,500 | 230,492 | 98.71 | | 10,000 | Before | 2,549,476 | 1,577,180 | 61.86 | | 10,000 | After | 2,546,468 | 1,574,172 | 61.82 |	6 years ago
Luke Imhoff	48b1de4617	Buffer formatted params instead of %TokenBalance{} ... ```elixir test "memory" do TokenBalance.Supervisor.Case.start_supervised!(json_rpc_named_arguments: []) pid = Process.whereis(TokenBalance.Fetcher) 10_000 |> insert_list(:token_balance) |> TokenBalance.Fetcher.async_fetch() Process.info(pid, :memory) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> IO.inspect(label: "AFTER GC") end ``` | Token Balances | Garbage Collection | `%TokenBalance{}`s (bytes) | `%{token_contract_address_hash: Hash.to_string(token_contract_address_hash), address_hash: Hash.to_string(address_hash), block_number: block_number}` (bytes) | Map / Struct (%) | |----------------|--------------------|----------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------------| | 1,000 | Before | 2,549,476 | 233,500 | 9.16 | | 1,000 | After | 2,546,468 | 233,500 | 9.17 | | 10,000 | Before | 19,903,172 | 2,549,476 | 12.81 | | 10,000 | After | 19,900,260 | 2,546,468 | 12.80 |	6 years ago
Luke Imhoff	753319de4a	Store ranges as tuples in Sequence :queue ... Found a memory savings for `Indexer.Sequence`: store Ranges as tuples in the `:queue` as the tuples take up less space than than the `%Range{}` struct, but only after garbage-collection. The incoming and outgoing format is still Ranges, so this is a non-breaking implementation detail only. ```elixir test "memory usage" do {:ok, pid} = Sequence.start_link(ranges: [max_block_number..0], step: -10) Process.info(pid, :memory) |> IO.inspect(label: "BEFORE GC") :erlang.garbage_collect(pid) Process.info(pid, :memory) |> IO.inspect(label: "AFTER GC") end ``` | max_block_number | Garbage Collection | Ranges (bytes) | Tuples (bytes) | Tuples / Ranges (%) | |------------------|--------------------|----------------|----------------|---------------------| | 5,000,000 | Before | 41,050,964 | 22,445,660 | 54.68 | | 5,000,000 | After | 34,386,756 | 23,879,996 | 69.45 | | 10,000,000 | Before | 80,889,772 | 47,928,116 | 59.24 | | 10,000,000 | After | 71,303,316 | 49,516,492 | 69.44 |	6 years ago
Sebastian Abondano	f58bf1586f	Log InternalTransaction.Fetcher errors as error ... Why: * Currently errors when importing internal transactions are being logged with 'debug' level. Making them 'error' level would make sure we notice them quicker and that they're prioritized accordingly. * Issue link: n/a	6 years ago
Amanda Sposito	0ed9020a96	Alter 'Realtime::Fetcher' and 'Catchup::Fetcher' to specify the broadcast type. ... The Token Counter needs to listen to both :realtime and :catchup events so it can keep up to date with the counting, but the 'Live Update' in the front-end doesn't need to hear the catchup events.	6 years ago
Luke Imhoff	1918040eeb	Explorer.Chain.Import.Runner behaviour	6 years ago
Luke Imhoff	59c7c38921	Extract Explorer.Chain.Import.Address.TokenBalances ... Standardized on :address_token_balances instead of :token_balances for options.	6 years ago
Luke Imhoff	991a2fe672	Extract Explorer.Chain.Import.Transactions	6 years ago
Luke Imhoff	22cb89058a	Extract Explorer.Chain.Import.Tokens	6 years ago
Luke Imhoff	14aa12e9bd	Extract Explorer.Chain.Import.TokenTransfers	6 years ago
Luke Imhoff	753fbd8bd5	Remove obsolete receipts	6 years ago
Luke Imhoff	5ffb54129f	Extract Explorer.Chain.Import.Logs	6 years ago
Luke Imhoff	4842462667	Extract Explorer.Chain.Import.Block.SecondDegreeRelations	6 years ago
Luke Imhoff	56f0365239	Extract Explorer.Chain.Import.Blocks	6 years ago
Luke Imhoff	03557f4200	Extract Explorer.Chain.Address.CoinBalances ... Standardize on `:address_coin_balances` instead of `:balances` for option names.	6 years ago
Nam Chu Hoai	d17917ce30	Allow unsecure WebSocket connections ... Some setup like ours will have the WebSocket connection in a private network to a local node and thus do not require a secure connection. This change allows to specify ws:// instead of wss:// endpoints as well as reading from the ENV var for the geth config of the indexer Co-authored-by: Luke Imhoff <luke.imhoff@dockyard.com>	6 years ago
Luke Imhoff	147a7e14fb	Don't do async imports with empty lists	6 years ago
Luke Imhoff	2e3627d536	Don't run imports for empty params	6 years ago
Andrew Cravenho	9892d1bd17	Add env variables to the production parity files	6 years ago
Luke Imhoff	bf7e8f11c5	Unsupport -> Unsupported	6 years ago
Luke Imhoff	9264a92db3	remaning -> remaining	6 years ago
Luke Imhoff	c60ae9467f	upto -> up to	6 years ago
Luke Imhoff	088589b11a	transaactions -> transactions	6 years ago
Luke Imhoff	65cbce71b4	registred -> registered	6 years ago
Luke Imhoff	1ba2e01bfa	lastest -> latest	6 years ago
Luke Imhoff	f87235c79a	strucutured -> structured	6 years ago
Alex Garibay	d7e38a223e	Missed token transfer cataloger (#807) ... * Rename token transfer params parser * Rename token transfer parsing function * Requeue blocks with incomplete token transfer indexing on start up * Add additional test for worker * Add missed token transfers to the front of the catchup queue * Address code review concerns * Remove unused alias * Add process names for uncataloged token transfer processes * Improved handling of when sequence isn't avaialbe to queue * Move Uncataloged.Supervisor name out of init and into start_link call Name is not a valid option to init as it is a GenServer.option, which means it is passed to the 3rd argument to Supervisor.start_link.	6 years ago
Felipe Renan	cc89b1f6db	Ignore burn address token balance for Tokens ERC-721 ... As we always get errors trying to fetch token balances when the address is the burn addres for Tokens ERC-721, we are gonna ignore them when we are processing the token balances from the token transfer.	6 years ago
Luke Imhoff	b8d9edf97d	Indexer.Block.Catchup.BoundIntervalSupervisor is proper supervisor ... Fixes #592 Ensure that Indexer.Block.Catchup.BoundIntervalSupervisor acts as a proper Supervisor by ensure it supports all Supervisor and :supervisor functions.	6 years ago
Luke Imhoff	2f7be3fb7d	OTP application descriptions ... The Application selector in Wobserver shows the descriptions, so it is useful to be set.	6 years ago
Sebastian Abondano	fdd0c84721	Realtime fetcher retries for unknown block number ... Why: * We want to add a retry mechanism, instead of waiting half a second for every block the realtime fetcher attempts to fetch and index. We had this half a second wait in place in an attempt to avoid "Unknown block number" errors. The half a second wait was effective at reducing these errors but it also meant we were awaiting unnecessarily for most fetches. Retrying after half a second, when Parity says it doesn't yet know of a given block number, makes more sense. * Issue link: n/a This change addresses the need by: * Editing `Indexer.Block.Realtime.Fetcher.fetch_and_import_block/3` to retry failed fetch/import attempts because of "Unknown block number" errors. It waits half a second before retrying and retries a maximum of three times.	6 years ago
Luke Imhoff	8e8b4b5508	Indexer.Block.Uncle.Fetcher ... * Use `Explorer.Chain.stream_unfetched_uncle_hashes` to initialize queue. * Have `Indexer.Block.Catchup.Fetcher` and `Indexer.Block.Realtime.Fetcher` queue uncles to fetch by calling `Indexer.Block.Uncle.Fetcher.async_fetch_blocks`.	6 years ago
Felipe Renan	e74bfb9d3b	Fix token balance log error ... We inverted the value of address_hash with the contract_address_hash. This commit fixes that.	6 years ago
Luke Imhoff	4b140c87df	Block in the consensus chain marked consensus true	6 years ago
Luke Imhoff	96d5aea311	Block.SecondDegreeRelation ... The database table block_second_degree_relations contains links between the `uncle_hash` of the uncle and the `nephew_hash` of the block that said `uncle_hash` was an uncle. It also has a `uncle_fetched_at` that is `null` until the `uncle_hash` is confirmed as fetched in the database. This does not contain such a fetcher.	6 years ago
Luke Imhoff	48b302f935	Return uncles from EtherumJSONRPC.Block.elixir_to_params	6 years ago
Stamates	ff5d4d0163	Wireup live update of pending transactions	6 years ago
Sebastian Abondano	8e3d80a6db	Realtime indexer waits half a second before fetch ... Why: * For Sokol, on staging, we noticed many errors like this one: ``` 00:45:05.627 application=indexer [error] realtime indexer failed to validate for block 4668478: [%{code: -32602, data: %{"blockNumber" => "0x473C3E", "hash" => "0x532e069e200be3e372da9680872826bb08caef91"}, message: "Unknown block number"}, %{code: -32602, data: %{"blockNumber" => "0x473C3E", "hash" => "0x95426f2bc716022fcf1def006dbc4bb81f5b5164"}, message: "Unknown block number"}, %{code: -32602, data: %{"blockNumber" => "0x473C3E", "hash" => "0xe797a1da01eb0f951e0e400f9343de9d17a06bac"}, message: "Unknown block number"}]. Block will be retried by catchup indexer. ``` * Issue link: https://github.com/poanetwork/blockscout/issues/808 This change addresses the need by: * Editing `Indexer.Block.Realtime.Fetcher.fetch_and_import_block/2` for it to wait half a second before starting to fetch and import. The rationale here is that we're hoping that half a second will be enough for Parity to sync and not return an "Unknown block number" error.	6 years ago
Sebastian Abondano	14b916f9e8	Realtime.Fetcher async fetching and no skips ... Why: * For the realtime fetcher to keep up on ETH we want it to fetch blocks asynchronously. We noticed that the realtime fetcher was falling behind on ETH mainnet. Currently the realtime fetcher fetches and imports one block at a time. This commit changes that, so that whenever we're notified of a new block number, through websockets, we start a new process to fetch and import each block asynchronously. * Also, whenever the websocket subscription, through which we're notified of new block numbers, skips block numbers, we want to make sure the skipped blocks are also fetched and imported. * Issue link: n/a This change addresses the need by: * Adding a new `Task.Supervisor` called `Indexer.Block.Realtime.TaskSupervisor` under the `Indexer.Block.Realtime.Supervisor` supervisor. This supervisor was created to supervise fetching and importing processes started by the realtime fetcher every time a new block number is received through the websocket subscription. * Editing `Indexer.Block.Realtime.Fetcher` to start a new process, under the new supervisor mentioned above, to fetch the latest block per the websocket `newHeads` subscription. It also determines if any block numbers have been skipped and starts individual processes to fetch and import the skipped blocks also.	6 years ago
Igor Florian	52e861fa73	Include and Log token_balances error attribute	6 years ago
Igor Florian	99012b2c4c	Add a new log file for token_balances fetching errors	6 years ago
Lucas Narciso	d8b22f6988	Handle large strings when fetching tokens	6 years ago