Manager node is responsible to orchestrate Kertish-dos farm. Default bind endpoint port is :9400

Manager node keep the index of files in Redis dss and cluster information in mongo db. Locking-center will use to keep index and cluster stability.

Should be started with parameters that are set as environment variables

• BIND_ADDRESS (optional) : Service binding address. Ex: 127.0.0.1:9400 Default: :9400

Head node will access the service using http://127.0.0.1:9400/client/manager and Data nodes will access the service using http://127.0.0.1:9400/client/node

• MONGO_CONN (mandatory) : Mongo DB endpoint. Ex: mongodb://admin:[email protected]:27017

Cluster and Data node setup and information will be kept in Mongo DB.

• MONGO_DATABASE (optional) : Mongo DB name. Default: kertish-dos

• MONGO_TRANSACTION (optional) : Set true if you have a Mongo DB Cluster setup

• REDIS_CONN (mandatory) : Redis dss. Ex: 127.0.0.1:6379

Will be used to index file information.

• REDIS_PASSWORD (optional) : Redis password for secure redis dss

• REDIS_CLUSTER_MODE (optional) : Redis cluster mode activation Ex: true

• LOCKING_CENTER (mandatory) : Locking-Center Server. Ex: 127.0.0.1:22119

Will be used to synchronize accesses between services

• HEALTH_CHECK_INTERVAL (optional) : Frequency of checking data-node(s) accessibility. default value is 10 seconds.

• GET is used to sync cluster/clusters, list cluster/clusters and nodes and find the cluster information for file.

• X-Action defines the behaviour of get request. Values: sync or repair or health or move or balance or clusters or find

• 422: Required Request Headers are not valid or absent

Sync action is to trigger the sync operation on cluster/clusters.

• X-Options header is used to point the cluster or omit it to run the sync operation in all clusters.

• 404: Not found
• 500: Operational failures
• 202: Accepted to finish in future

All failed responses comes with error json. Ex:

{
  "code": 100,
  "message": "clusters are not available for sync"
}

Repair action is to trigger the consistency and integrity repair operation on cluster/clusters.

• X-Options header is used to point the repair type. Omit for full repair, structure for only metadata structure repair or integrity for only data-node and metadata integrity repair.

• 404: Not found
• 423: Locked for another repair operation
• 500: Operational failures
• 202: Accepted to finish in future

All failed responses comes with error json. Ex:

{
  "code": 105,
  "message": "clusters are not available for repair"
}

Health action is to get the data-node accessibility report.

• 500: Operational failures
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 140,
  "message": "clusters are not available for repair"
}

Sample report output:

{
  "[clusterId]": [
    {
      "nodeId": "2f194705b3b6292e84dc71dbb0185a9c",
      "address": "172.20.1.40:9430",
      "master": true,
      "quality": 0
    }
  ]
}

Possible quality values:

• -2 DNS Error, unable to resolve
• -1 Paralysed or Readonly Data-Node/Cluster
• 0 or higher Response Time in ms

Move action is to move one cluster content to other one. Target cluster should have enough space for move operation.

• X-Options header is used to point the source and target clusters for move operation. sourceClusterId,targetClusterId

• 404: Not found
• 422: Invalid Headers for operation
• 500: Operational failures
• 503: Service Unavailable
• 507: Insufficient Space
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 130,
  "message": "cluster is not available for cluster wide actions"
}

Balance action is to balance data weight between clusters.

• X-Options header is used to point specific clusters to balance between. clusterId,clusterId,...

• 404: Not found
• 422: Invalid Headers for operation
• 500: Operational failures
• 503: Service Unavailable
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 135,
  "message": "cluster is not available for cluster wide actions"
}

Clusters action is to get the information about the cluster/clusters and related data nodes.

• X-Options header is used to point the cluster or omit it to get al clusters information.

• 404: Not found
• 500: Operational failures
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 110,
  "message": "cluster is already exists"
}

Find action is to get the clusterId and data node of a file.

• X-Options header is used to point the fileId. FileId is a sha512 encoded hex string. Ex: e5c0adae0f05cf60f7e34b45bd44249f42627b1f3b1b453ae45e106adbfdfbdb

• Successful response contains X-Cluster-Id and X-Address for search result.

• 404: Not found
• 500: Operational failures
• 503: Not available to complete the operation (Readonly, Offline or Paralysed cluster/node)
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 120,
  "message": "cluster is already exists"
}

• POST is used to create cluster, register node, take snapshot, make reservation, create read and delete maps.

• X-Action defines the behaviour of post request. Values: register or snapshot or reserve or readMap or createMap or deleteMap

• 422: Required Request Headers are not valid or absent

Register action will create cluster and/or register node.

• X-Options header contains the registration or addition details. Format is clusterId=[DataNodeBındingAddress]:[DataNodeBındingPort],[DataNodeBındingAddress]:[DataNodeBındingPort],...

if you omit the clusterId, request will be accepted as to create new cluster with the given data nodes.

to create new cluster: Ex: 127.0.0.1:9430,127.0.0.1:9431

to add new data nodes to the existence cluster: Ex: 8f0e2bc02811f346d6cbb542c92d118d=127.0.0.1:9430,127.0.0.1:9431

• 400: Operational failure
• 409: Cluster is already created/Data Node is already registered
• 422: Required Request Headers are not valid or absent
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 200,
  "message": "cluster is already exists"
}

Successful request response sample

{
  "clusterId": "8f0e2bc02811f346d6cbb542c92d118d",
  "size": 1073741824,
  "used": 0,
  "nodes": [
    {
      "nodeId": "7a758a149e4453b20a40b35f83f3a0e4",
      "address": "127.0.0.1:9430",
      "master": true,
      "quality": 0
    }
  ],
  "reservations": []
}

Snapshot action will create a snapshot state for the cluster.

• X-Options header is used to send the cluster id to take the snapshot for.

• 400: Operational failure
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 205,
  "message": "cluster is already exists"
}

Reserve action is to reserve data space on data nodes to guaranteed that files can be stored.

• X-Size header uint64 value for the required space size.

• 400: Operational failures
• 422: Required Request Headers are not valid or absent
• 503: Not available for reservation (Readonly, Offline or Paralysed cluster/node)
• 507: Insufficient space
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 210,
  "message": "cluster is already exists"
}

Successful request response sample

{
  "reservationId": "e813bfe9-50fd-4a13-992d-10995326210f",
  "clusters": [
    {
      "clusterId": "8f0e2bc02811f346d6cbb542c92d118d",
      "address": "127.0.0.1:9430",
      "chunk": {
        "sequence": 1,
        "index": 0,
        "size": 10247680
      } 
    }
  ]
}

Creates the cluster access map for the specified files.

• X-Options header holds the file hex id list with , separated. Ex: e5c0adae0f05cf60f7e34b45bd44249f42627b1f3b1b453ae45e106adbfdfbdb,45bd44249f42627b1f3b1b453ae45e106adbfdfbdba5c0adae0f05cf60f7e34b

• 400: Operational failures
• 422: Required Request Headers are not valid or absent
• 503: Not available for reservation (Readonly, Offline or Paralysed cluster/node)
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 220,
  "message": "cluster is already exists"
}

Successful request response sample

{
  "e5c0adae0f05cf60f7e34b45bd44249f42627b1f3b1b453ae45e106adbfdfbdb": "127.0.0.1:9430",
  "45bd44249f42627b1f3b1b453ae45e106adbfdfbdba5c0adae0f05cf60f7e34b": "127.0.0.1:9431"
}

• DELETE is used to delete cluster, unregister node, delete snapshot, discard or commit reservation.

• X-Action defines the behaviour of delete request. Values: unregister or snapshot or commit or discard

• 422: Required Request Headers are not valid or absent

Unregister action will delete cluster and/or unregister node.

• X-Options header contains the unregistering or deletion details. Format is [type],[clusterId/nodeId]

type values are c for cluster and n for node. You should push the related id for the operation Ex: c,8f0e2bc02811f346d6cbb542c92d118d

• 404: Cluster/Node not found
• 422: Required Request Headers are not valid or absent
• 423: (only node deletion) Last node cannot be deleted
• 500: Operational failure
• 200: Successful

All failed responses comes with error json. code can be 300 for cluster or 350 for node operation Ex:

{
  "code": 300,
  "message": "cluster is already exists"
}

Change state action will change the cluster current state between 3 states (Online, Readonly, Offline).

• X-Options header contains the action delegation for each or all clusters. Format is [clusterId]=[state],[clusterId]=[state],...

states are:

• 0: Online
• 1: Readonly
• -1: Offline

• 404: Cluster/Node not found
• 422: Required Request Headers are not valid or absent
• 500: Operational failure
• 200: Successful

All failed responses comes with error json. code can be 300 Ex:

{
  "code": 300,
  "message": "clusters state change options are insufficient"
}

Snapshot action will delete the snapshot for the cluster pointed with snapshot index.

• X-Options header contains the clusterId and snapshot index with = separator. Ex: clusterId=snapshotIndex

• 422: Required Request Headers are not valid or absent
• 500: Operational failure
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 380,
  "message": "cluster is already exists"
}

Commit action is to commit the reserved disk space for the further process.

• X-Reservation-Id header for registration id.
• X-Options for committing details. It has a special format clusterId=size,clusterId=size,...

clusterId can repeat and the size will grow base on clusterId matching. size must be uint64.

when reservation is committed once with the clusterId-size map, unused space will be added to the cluster total space.

• 422: Required Request Headers are not valid or absent
• 500: Operational failures
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 360,
  "message": "cluster is already exists"
}

Discard action is to discard the reserved disk space in all related clusters.

• X-Reservation-Id header for registration id.

when reservation is discarded once, unused space will be added to the cluster total space.

• 400: Operational failures
• 422: Required Request Headers are not valid or absent
• 200: Successful

All failed responses comes with error json. Ex:

{
  "code": 370,
  "message": "cluster is already exists"
}