5.2. Die Konfiguration

In diesem Abschnitt werden wir auf die Konfiguration von Iroha eingehen, ausgehend von ‚example/config.sample‘.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
{
  "block_store_path": "/tmp/block_store/",
  "torii_port": 50051,
  "internal_port": 10001,
  "pg_opt": "host=localhost port=5432 user=postgres password=mysecretpassword",
  "max_proposal_size": 10,
  "proposal_delay": 5000,
  "vote_delay": 5000,
  "mst_enable" : false,
  "mst_expiration_time" : 1440,
  "max_rounds_delay": 3000,
  "stale_stream_max_rounds": 2
}

Die Konfigurationsdatei ist eine valide JSON-Struktur. Lass‘ uns diese Zeile für Zeile ansehen und verstehen, was jeder Parameter bedeutet.

5.2.1. Deployment-spezifische Parameter

  • ‚block_store_path‘ setzt den Pfad an dem Blöcke gespeichert werden
  • ‚torii_port‘ setzt den Port für externe Kommunikation. Queries und Transaktionen werden hierhin gesendet.
  • ‚internal_port‘ setzt den Port für interne Kommunikation: Ordering service, Konsensus und block loader
  • ‚pg_opt‘ setzt Zugangsdaten für PostgreSQL: Hostname, Port, Nutzername und Passwort
  • log is an optional parameter controlling log output verbosity and format (see below).

5.2.2. Umgebungsspezifische Parameter

  • ‚max_proposal_size‘ ist die maximale Anzahl an Transaktionen, die in einem Proposal, und darausfolgend in einem Block sein können. Die Änderung dieses Wertes definiert auch die Größe eines potentiellen Blockes. Als Startwert empfiehlt sich ‚10‘. Sollten eine große Menge an Transaktionen pro Sekunden zu erwarten sein, empfiehlt es sich diesen Wert zu erhöhen.

  • proposal_delay is a timeout in milliseconds that a peer waits a response from the orderding service with a proposal.

  • ‚vote_delay‘ ist die Wartezeit in Millisekunden bevor ein Votum zum nächsten Peer gesendet wird. Ein optimaler Wert ist stark abhängig von der Anzahl von Iroha-Peers im Netzwerk (eine größere Anzahl von Nodes erfordert einen größeren ‚vote_delay‘). Wir empfehlen einen Startwert von 100 bis 1000 ms.

  • mst_enable enables or disables multisignature transaction network transport in Iroha. Note that MST engine always works for any peer even when the flag is set to false. The flag only allows sharing information about MST transactions among the peers.

  • mst_expiration_time is an optional parameter specifying the time period in which a not fully signed transaction (or a batch) is considered expired (in minutes). The default value is 1440.

  • max_rounds_delay is an optional parameter specifying the maximum delay between two consensus rounds (in milliseconds). The default value is 3000. When Iroha is idle, it gradually increases the delay to reduce CPU, network and logging load. However too long delay may be unwanted when first transactions arrive after a long idle time. This parameter allows users to find an optimal value in a tradeoff between resource consumption and the delay of getting back to work after an idle period.

  • stale_stream_max_rounds is an optional parameter specifying the maximum amount of rounds to keep an open status stream while no status update is reported. The default value is 2. Increasing this value reduces the amount of times a client must reconnect to track a transaction if for some reason it is not updated with new rounds. However large values increase the average number of connected clients during each round.

  • "initial_peers is an optional parameter specifying list of peers a node will use after startup instead of peers from genesis block. It could be useful when you add a new node to the network where the most of initial peers may become malicious. Peers list should be provided as a JSON array:

    "initial_peers" : [{"address":"127.0.0.1:10001", "public_key": "bddd58404d1315e0eb27902c5d7c8eb0602c16238f005773df406bc191308929"}]

5.2.3. Logging

In Iroha logging can be adjusted as granularly as you want. Each component has its own logging configuration with properties inherited from its parent, able to be overridden through config file. This means all the component loggers are organized in a tree with a single root. The relevant section of the configuration file contains the overriding values:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
"log": {
  "level": "info",
  "patterns": {
    "debug": "don't panic, it's %v.",
    "error": "MAMA MIA! %v!!!"
  },
  "children": {
    "KeysManager": {
      "level": "trace"
    },
    "Irohad": {
      "children": {
        "Storage": {
          "level": "trace",
          "patterns": {
            "debug": "thread %t: %v."
          }
        }
      }
    }
  }
}

Every part of this config section is optional.

  • level sets the verbosity. Available values are (in decreasing verbosity order):
    • trace - print everything
    • debug
    • info
    • warning
    • error
    • critical - print only critical messages
  • patterns controls the formatting of each log string for different verbosity levels. Each value overrides the less verbose levels too. So in the example above, the „don’t panic“ pattern also applies to info and warning levels, and the trace level pattern is the only one that is not initialized in the config (it will be set to default hardcoded value).
  • children describes the overrides of child nodes. The keys are the names of the components, and the values have the same syntax and semantics as the root log configuration.