BigQuery失败的加载作业 - 无法访问GCS文件 [英] BigQuery failed load job - was unable to access a GCS file

问题描述

昨天，我们在BigQuery的一个计划加载作业中遇到了一个错误。我们以前从未见过这个错误。此外，不久之后重新开展工作（没有任何改变）工作得很好。因此，无法重现。

We encountered an error in one of our scheduled load jobs for BigQuery yesterday. We've never seen this error before. In addition, re-running the job soon afterwards (without any changes) worked fine. Thus, cannot reproduce.

我们正在从GCS加载文件。以下是BigQuery / GCS返回的错误：
$ b

We're loading the files in from GCS. Here's the error that BigQuery/GCS came back with:

访问被拒绝：Google存储文件[已删除文件名]：拒绝访问

Access Denied: Google Storage File [file name removed]: Access Denied

只有〜50个源文件中的一个文件不可访问。
$ b

Just that one file out of ~50 source uri's was not accessible. The job id of failed job:

job_TFuRwMrEWVAMD-PNuc29Yx_PXkk

重新运行作业的作业ID，在当天晚些时候再次正常工作（不做任何更改）：

The job id of the re-run job later in the day which worked fine again (no changes made):

job_Pt9usLRO9elxIBlmuOwCGB-drxI

job_Pt9usLRO9elxIBlmuOwCGB-drxI

任何人都有一个想法，为什么会发生这种情况？

Anyone have an idea why this happened?

推荐答案

我在BigQuery导入工作人员中查找了作业日志...从我可以告诉我们何时试图获得其中一个Cloud Storage文件的文件大小时，我们遇到了ACCESS_DENIED错误。你确定ACL在第二份工作运行之前没有改变吗？这可能是Cloud Storage中的暂时故障。

I looked up the job logs in the BigQuery import workers... from what I can tell when we tried to get the file size of one of Cloud Storage files we hit an ACCESS_DENIED error. Are you certain that the ACL did not change before the second job was run? It may have been a temporary glitch in Cloud Storage.

这篇关于BigQuery失败的加载作业 - 无法访问GCS文件的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！