5.2. 설정

이 섹션에서는 Iroha를 설정하는 방법에 대해 설명합니다. 우선 ``example/config.sample``을 봅시다.

 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
}

위에서 확인할 수 있듯이, 설정 파일은 json 파일입니다. 한 줄 한 줄 확인해보면서 모든 파라미터들이 무엇을 의미하는지 알아봅시다.

5.2.1. Deployment-specific parameters

  • ``block_store_path``은 블록을 저장할 폴더의 경로를 설정합니다.
  • ``torii_port``는 외부 통신을 위한 포트를 설정합니다. 쿼리와 트랜젝션은 여기를 통해 전송됩니다.
  • ``internal_port``는 오더링 서비스(Ordering Service)나 합의(Consensus), 블록 로더(Block Loader)와 같은 내부 통신을 위한 포트를 설정합니다.
  • ``pg_opt``는 PostgreSQL의 자격증명(Credentials)를 설정하는 데 사용됩니다. 호스트 이름, 포트, 유저 이름, 패스워드를 설정할 수 있습니다.
  • log is an optional parameter controlling log output verbosity and format (see below).

5.2.2. 환경(Environment) 관련 파라미터

  • max_proposal_size is the maximum amount of transactions that can be in one proposal, and as a result in a single block as well. So, by changing this value you define the size of potential block. For a starter you can stick to 10. However, we recommend to increase this number if you have a lot of transactions per second.

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

  • vote_delay is a waiting time in milliseconds before sending vote to the next peer. Optimal value depends heavily on the amount of Iroha peers in the network (higher amount of nodes requires longer vote_delay). We recommend to start with 100-1000 milliseconds.

  • 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.