[
https://issues.apache.org/jira/browse/BEAM-12447?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17549344#comment-17549344
]
Danny McCormick commented on BEAM-12447:
----------------------------------------
This issue has been migrated to https://github.com/apache/beam/issues/20947
> SdkContainerImageBuilder should take machineType configs
> --------------------------------------------------------
>
> Key: BEAM-12447
> URL: https://issues.apache.org/jira/browse/BEAM-12447
> Project: Beam
> Issue Type: Improvement
> Components: runner-dataflow, sdk-py-core
> Reporter: Yichi Zhang
> Priority: P3
> Time Spent: 4h 10m
> Remaining Estimate: 0h
>
> When prebuilding the python SDK container harness image with
> SdkContainerImageBuilder on google cloud build, it can run out of memory and
> fail the build with docker exit code 137. The default machineType is only
> n1-standard-1 and it is unlikely sufficient to build large images.
>
> The SdkContainerImageBuilder should take additional machineType configs to
> circumvent the docker OOM issue.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
