[
https://issues.apache.org/jira/browse/TEZ-1213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14039193#comment-14039193
]
Siddharth Seth commented on TEZ-1213:
-------------------------------------
bq. some fields in TezRuntimeFrameworkConfigs do not use TEZ_RUNTIME_PREFIX
Fixed
bq. For io.file.buffer.size, is there a hadoop-common defined var which we can
re-use?
No. Everything references this parameter directly in hadoop-trunk.
bq. There are some inconsistencies like
TEZ_RESOURCE_CALCULATOR_PROCESS_TREE_CLASS which uses tez runtime prefix but
the field name is unchanged. What var is not prefixed with tez.runtime apart
from credentials?
Fixed. Other than credentials, everything should have the prefix.
bq. Spurious comments like // KKK
Removed
bq. Utility code to generate statements to add every element to the keyset
Will just remove this. That was used for generating the list of entities for
the keySet. Not meant for user consumption.
bq. String basePath =
conf.get(TezRuntimeFrameworkConfigs.TASK_LOCAL_RESOURCE_DIR, ".");
Let me see what can be done with this.
bq. It is surprising that there are hardly any changes in
registerMRToRuntimeKeyTranslation() with respect to the re-factored names?
Very few names are actually refactored. Most of the changes are to use
TEZ_PREFIX in the key value, rather than explicitly having "tez.runtime." in
every key.
> Fix parameter naming in TezJobConfig
> ------------------------------------
>
> Key: TEZ-1213
> URL: https://issues.apache.org/jira/browse/TEZ-1213
> Project: Apache Tez
> Issue Type: Improvement
> Reporter: Siddharth Seth
> Assignee: Siddharth Seth
> Attachments: TEZ-1213.1.txt
>
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
