cluster_v1_remoteexecnodes.yaml

apiVersion: cluster.ytsaurus.tech/v1
kind: RemoteExecNodes
metadata:
  name: remote-exec-nodes
spec:
  remoteClusterSpec:
    name: remote-ytsaurus

  # FIXME: Move cluster options into RemoteYtsaurus.
  coreImage: ghcr.io/ytsaurus/ytsaurus:dev-23.2-relwithdebinfo

  jobImage: mirror.gcr.io/library/python:3.12-slim

  # configOverrides:
  #  name: ytsaurus-config-overrides

  useIpv4: true
  # useIpv6: true

  # hostNetwork: true

  name: remote
  instanceCount: 1

  loggers: &loggers
    - name: debug
      compression: zstd
      minLogLevel: debug
      writerType: file
      rotationPolicy: &rotationPolicy
        maxTotalSizeToKeep: 1073741824 # 1GiB
        rotationPeriodMilliseconds: 900000 # 15Min
        maxSegmentCountToKeep: 1000
      categoriesFilter:
        type: exclude
        values: [ "Bus", "Concurrency" ]
    - name: info
      minLogLevel: info
      writerType: file
      rotationPolicy: *rotationPolicy
    - name: info-stderr
      minLogLevel: info
      writerType: stderr

  jobProxyLoggers:
    - name: debug
      compression: zstd
      minLogLevel: debug
      writerType: file
      useTimestampSuffix: true
      rotationPolicy: &rotationPolicyJobs
        maxTotalSizeToKeep: 104857600  # 100Mi
        rotationPeriodMilliseconds: 900000  # 15Min
      categoriesFilter:
        type: exclude
        values: [ "Bus", "Concurrency" ]
    - name: info
      minLogLevel: info
      writerType: file
      rotationPolicy: *rotationPolicyJobs
    - name: error
      minLogLevel: error
      writerType: stderr

  resources:
    # Allocate resources for exec node container
    requests:
      cpu: 1
      memory: 1Gi
    limits:
      cpu: 10
      memory: 10Gi

  locations:
    - locationType: Logs
      path: /yt/exec-node-logs
    - locationType: ChunkCache
      path: /yt/node-data/chunk-cache
    - locationType: Slots
      path: /yt/node-data/slots
    - locationType: ImageCache
      path: /yt/node-data/image-cache

  volumeMounts:
    - name: exec-node-logs
      mountPath: /yt/exec-node-logs
    - name: node-data
      mountPath: /yt/node-data
      # mountPropagation: Bidirectional  # Enable for tmpfs in jobs

  volumeClaimTemplates:
    - metadata:
        name: node-data
      spec:
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: 20Gi
    - metadata:
        name: exec-node-logs
      spec: &logsVolumeSpec
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: 10Gi

  # privileged: true  # Enable for tmpfs in jobs

  jobResources:
    # Allocate resources for jobs container
    requests:
      cpu: 4
      memory: 4Gi
    limits:
      cpu: 10
      memory: 10Gi

  jobEnvironment:
    # Add CRI containerd sidecar
    cri:
      apiRetryTimeoutSeconds: 180