[
https://issues.apache.org/jira/browse/TEZ-3856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16217812#comment-16217812
]
Zhiyuan Yang commented on TEZ-3856:
-----------------------------------
New patch is functionally working and should be able to meet the requirement in
your case. I agree with you that we should store the additional counters in
vertex. Two comments regarding to API:
1. InputInitializerContext.java
{code}
TezCounters getCounters();
{code}
We don't need expose counter to user level code if input initializer just need
update the counter. An api that allow input initializer to update the counter
should be good enough.
2. Vertex.java
{code}
TezCounters getCounters();
{code}
This one potentially can cause confusion since Task.getCounter is already used
and means different thing. As like last comments, an api that allow update on
custom counter is good enough. By doing this we will be able to avoid undesired
counter update from user code, like, conflicts between user's counter and
internal counters.
> API to access counters in InputInitializerContext
> -------------------------------------------------
>
> Key: TEZ-3856
> URL: https://issues.apache.org/jira/browse/TEZ-3856
> Project: Apache Tez
> Issue Type: Bug
> Affects Versions: 0.9.1
> Reporter: Prasanth Jayachandran
> Assignee: Prasanth Jayachandran
> Attachments: TEZ-3856.1.patch, TEZ-3856.2.patch, TEZ-3856.2.patch
>
>
> Hive would like to publish some counters related to input splits during split
> generation. Tez doesn't expose TezCounters via InputIntializerContext. This
> ticket is to expose TezCounters via InputInitializerContext so that counters
> can be accessed during split generation.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
