0

我经常有一个被消息冻结的容器

[
    {
        "level": "info",
        "msg": "No network configured. Not Joining one.",
        "time": "2020-02-28T11:46:02Z"
    },
    {
        "action": "esvector_startup",
        "level": "info",
        "maxWaitTime": 120000000000,
        "msg": "waiting for es vector to start up (maximum 2m0s)",
        "time": "2020-02-28T11:46:02Z"
    },
    {
        "action": "restapi_management",
        "level": "info",
        "msg": "Serving weaviate at http://[::]:8080",
        "time": "2020-02-28T11:46:02Z"
    },
    {
        "level": "warn",
        "ts": "2020-02-29T14:39:54.813Z",
        "caller": "clientv3/retry_interceptor.go:61",
        "msg": "retrying of unary invoker failed",
        "target": "endpoint://client-5b29a880-9d71-462e-b7ac-a573d5e2e6e5/etcd:2379",
        "attempt": 0,
        "error": "rpc error: code = Unavailable desc = etcdserver: request timed out"
    }
]

在 ETCD-0 我有以下日志: **在 ETCD-0 我有以下日志:

2020-02-29 14:39:56.955415 I | raft: 4871a93f1a47265c [term: 316] ignored a MsgHeartbeatResp message with lower term from 5b97572db0dcba3a [term: 311]
2020-02-29 14:39:57.805445 I | raft: 4871a93f1a47265c is starting a new election at term 316
2020-02-29 14:39:57.805480 I | raft: 4871a93f1a47265c became candidate at term 317
2020-02-29 14:39:57.805493 I | raft: 4871a93f1a47265c received MsgVoteResp from 4871a93f1a47265c at term 317
2020-02-29 14:39:57.805503 I | raft: 4871a93f1a47265c [logterm: 311, index: 70906] sent MsgVote request to 5b97572db0dcba3a at term 317
2020-02-29 14:39:57.845489 I | raft: 4871a93f1a47265c received MsgVoteResp from 5b97572db0dcba3a at term 317
2020-02-29 14:39:57.845522 I | raft: 4871a93f1a47265c [quorum:2] has received 2 MsgVoteResp votes and 0 vote rejections
2020-02-29 14:39:57.845541 I | raft: 4871a93f1a47265c became leader at term 317
2020-02-29 14:39:57.845551 I | raft: raft.node: 4871a93f1a47265c elected leader 4871a93f1a47265c at term 317
2020-02-29 15:37:06.688476 N | compactor: Starting auto-compaction at revision 68399 (retention: 4h0m0s)

在我的 ETCD-1 上,我有以下日志

2020-02-29 13:37:06.663962 I | mvcc: finished scheduled compaction at 66935 (took 859.28µs)
2020-02-29 14:37:06.685573 I | mvcc: store.index: compact 67655
2020-02-29 14:37:06.688036 I | mvcc: finished scheduled compaction at 67655 (took 1.405431ms)
2020-02-29 14:39:51.509681 I | raft: 5b97572db0dcba3a [logterm: 311, index: 70906, vote: 4871a93f1a47265c] ignored MsgVote from 4871a93f1a47265c [logterm: 311, index: 70906] at term 311: lease is not expired (remaining ticks: 10)
2020-02-29 14:39:52.811383 I | raft: 5b97572db0dcba3a [logterm: 311, index: 70906, vote: 4871a93f1a47265c] ignored MsgVote from 4871a93f1a47265c [logterm: 311, index: 70906] at term 311: lease is not expired (remaining ticks: 10)
2020-02-29 14:39:54.610822 I | raft: 5b97572db0dcba3a [logterm: 311, index: 70906, vote: 4871a93f1a47265c] ignored MsgVote from 4871a93f1a47265c [logterm: 311, index: 70906] at term 311: lease is not expired (remaining ticks: 10)
2020-02-29 14:39:55.610694 I | raft: 5b97572db0dcba3a [logterm: 311, index: 70906, vote: 4871a93f1a47265c] ignored MsgVote from 4871a93f1a47265c [logterm: 311, index: 70906] at term 311: lease is not expired (remaining ticks: 10)

重新启动容器后,一切正常。

我已经在 K8S 上部署了 Weaviate,并为 ETCD 提供了以下配置:

# Etcd
#
# Weaviate stores critical configuration where strong consistency is required
# in etcd.
etcd:
  fullnameOverride: etcd
  envVarsConfigMap: 'etcd-config'
  statefulset:
    replicaCount: 2
  ##
  auth:
    rbac:
      enabled: false
    client:
      ## Switch to encrypt client communication using TLS certificates
      secureTransport: false
      ## Switch to automatically create the TLS certificates
      useAutoTLS: false
      enableAuthentication: false
    peer:
      ## Switch to encrypt client communication using TLS certificates
      secureTransport: true
      ## Switch to automatically create the TLS certificates
      useAutoTLS: true
      ## Switch to enable host authentication using TLS certificates. Requires existing secret.
      enableAuthentication: false
  metrics:
    enabled: true
    podAnnotations:
      prometheus.io/scrape: 'true'
    prometheus.io/port: '2379'
  disasterRecovery:
    # If you set `enabled: true` you need to make sure that an NFS provisioner
    # runs in your cluster! See
    # https://github.com/bitnami/charts/tree/master/bitnami/etcd#disaster-recovery
    # Defaults to 'false' so the chart works without an NFS provisioner.
    # However, 'enabled: true' is strongly recommended!
    enabled: false
    cronjob:
      schedule: '*/30 * * * *'
      historyLimit: 1
      podAnnotations: {}
    pvc:
      size: 2Gi
      storageClassName: default
  startFromSnapshot:
    enabled: false
    ## Existingn PVC containing the etcd snapshot
    ##
    # existingClaim
    ## Snapshot filename
    ##
    # snapshotFilename
    #

我的配置有什么问题导致冻结 Weaviate 容器吗?

谢谢!

4

0 回答 0