[
https://issues.apache.org/jira/browse/LENS-743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15048568#comment-15048568
]
Amareshwari Sriramadasu commented on LENS-743:
----------------------------------------------
bq. Query handle will be same, all the state transitions will be same, just
that after FAILED, it'll move to QUEUED/RUNNING state.
This would add to confusion. We can associate Query to have attempts and each
attempt will be associated with all states. The id of attempt can be
<query_handle>_attempt<id>, if that makes it simple. Each attempt will have the
details as they are available for query now.
bq. > errormessage = Latest error msg or appended error messages . How should
we (if at all) show error messages from previous runs ? If the query succeeds
finally will we wipe this field empty ?
We should never wipe of old states.
> Query failure retries for transient errors
> ------------------------------------------
>
> Key: LENS-743
> URL: https://issues.apache.org/jira/browse/LENS-743
> Project: Apache Lens
> Issue Type: Improvement
> Components: server
> Reporter: Amareshwari Sriramadasu
> Assignee: Rajat Khandelwal
>
> There have to be retries for query failures for transient errors like network
> errors (Hive server not reachable/ Metastore not reachable/ DB not
> reachable). Retries should be available for each phase - submission,
> execution, updating status, fetching results and formatting.
> Right now, any such failure results in marking query as failed.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
