[CRON] Failed: apache/flink#42140 (release-1.10 - 91d96ab)

[Flink CI <fl...@foxmail.com>]

|  [

apache

/

flink

](https://travis-ci.org/apache/flink)  
---  
  
![branch icon](https://s3.amazonaws.com/travis-email-
assets/branch.png)[release-1.10](https://github.com/apache/flink/tree/release-1.10)  
  
|

![build hasfailed](https://s3.amazonaws.com/travis-email-assets/status-
failed.png)

[Build #42140 failed](https://travis-ci.org/apache/flink/builds/638048043)

![arrow to build time](https://s3.amazonaws.com/travis-email-assets/failure-
arrow.png)

![clock icon](https://s3.amazonaws.com/travis-email-assets/time.png)31 hrs, 43
mins, 53 secs  
  
---  
  
![Xintong Song
avatar](https://secure.gravatar.com/avatar/8e71b67227007b42bf32f16b86f428e9.jpg)Xintong
Song

[91d96ab CHANGESET
->](https://github.com/apache/flink/compare/7b6aa1a63f45742557efc5d0bef2db40ebcc683d...91d96abe5f42bd088a326870b4885d79611fccb5)  
  
[FLINK-15598][yarn] Do not overwrite configuration with cluster specification
in YarnClusterDescriptor#startAppMaster.  
  
ClusterSpecification is generated from configuration at the first place, in
AbstractContainerizedClusterClientFactory. It does not make sense to write the
values back.  
  
We are aware that ClusterSpecification#taskManagerMemoryMB might be updated in
YarnClusterDescriptor#validateClusterResources, which is before reaching
YarnClusterDescriptor#startAppMaster. However, this will only happen if the TM
process memory is configured smaller than Yarn minimum allocation. While using
the original process memory size may lead to waste, it should not cause any
serious problems or failures. On the other hand, overwriting the process
memory size may lead to failures due to memory configuration conflicts, of
which the reason could be implict to the users. Thus, we believe throwing a
warning on process memory less than Yarn minimum allocation but not
overwriting it would be enough.  
  
This closes #10863.