gcloud dataproc集群更新-如何删除导致命令成功运行的未决更新

时间:2019-02-22 19:40:21

标签: google-cloud-platform gcloud google-cloud-dataproc

在我的google dataproc集群中,我尝试通过停用一个datanode来动态调整集群的大小。进行集群更新,并增加工作人员数量,以将集群大小减小为1个主服务器和5个工作人员,而这两个小时之内尚未完成。

ERROR: (gcloud.dataproc.clusters.update) FAILED_PRECONDITION: Cannot update cluster 'my-cluster-dev' while it has other pending update operations.

我锁定了计算机,并以某种方式丢失了命令外壳,群集仍显示其处于更新状态,从而阻止了我执行任何操作。我现在已经花费了大约5个小时,但仍然无法执行任何命令或访问集群。

gcloud dataproc clusters describe my-cluster-dev

projectId: my-project-id-dev
status:
  state: UPDATING
  stateStartTime: '2019-02-22T12:30:02.084Z'
statusHistory:

任何人都可以帮助我解决导致群集陷入更新状态的进程,以便我可以在群集上执行其他任务吗?

2 个答案:

答案 0 :(得分:1)

What steps did you take in order to reduce number of nodes in your cluster? It's unclear what you mean by "dynamically resize the cluster by decommissioning one datanode". Are you saying you ran "gcloud dataproc clusters update my-cluster-dev --num-masters=1 --num-workers=5" when my-cluster-dev had 6 nodes and it got stuck into updating state or were there any additional operations that you performed?

答案 1 :(得分:0)

遇到了同样的问题。试图更新工作人员的数量,并陷入“更新”状态。像OP一样,手动停止/启动主节点解决了该问题。