[
https://issues.apache.org/jira/browse/YARN-3908?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14643188#comment-14643188
]
Vrushali C commented on YARN-3908:
----------------------------------
Hi Zhijie
Thanks.. that is a good point. But if we put the event timestamp first, we have
no way of querying for a particular event key unless we know the exact
timestamp. I think knowing the exact time is probably almost impossible.
Imagine that there is another event that occurs between the two kill events, so
it has a timestamp > kill1 and < kill2. Now we still have to fetch all those
and filter them out. So placing the timestamp first does not help in this case.
But if we have the event key first, the columns will be placed together and the
event timestamps will be stored in a chronological order (using the long.max -
ts value). So the first one being fetched for kill event would be the latest
for that event key.
thanks
Vrushali
> Bugs in HBaseTimelineWriterImpl
> -------------------------------
>
> Key: YARN-3908
> URL: https://issues.apache.org/jira/browse/YARN-3908
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: timelineserver
> Reporter: Zhijie Shen
> Assignee: Vrushali C
> Attachments: YARN-3908-YARN-2928.001.patch,
> YARN-3908-YARN-2928.002.patch, YARN-3908-YARN-2928.003.patch,
> YARN-3908-YARN-2928.004.patch, YARN-3908-YARN-2928.004.patch,
> YARN-3908-YARN-2928.005.patch
>
>
> 1. In HBaseTimelineWriterImpl, the info column family contains the basic
> fields of a timeline entity plus events. However, entity#info map is not
> stored at all.
> 2 event#timestamp is also not persisted.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
