[
https://issues.apache.org/jira/browse/HIVE-24753?focusedWorklogId=557782&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-557782
]
ASF GitHub Bot logged work on HIVE-24753:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 25/Feb/21 07:55
Start Date: 25/Feb/21 07:55
Worklog Time Spent: 10m
Work Description: pvargacl edited a comment on pull request #2017:
URL: https://github.com/apache/hive/pull/2017#issuecomment-785691173
I missing some features, should these part of this PR?
- shouldn't the lock type changed from exclusive to excl_write if we use
this kind of drop?
- shouldn't he Cleaner later delete the whole partition directory if it was
not recreated?
- I think you should add many test, around drop and recreate, what happens
if you recreate in different scenarios (while there is an old read still
running, when you overlap with the cleaner, when it was already cleaned up ...
- there will be a race condition in the cleaner and the create partition I
think, what happens if Cleaner starts to run, checks if the partition was
dropped, so decides it has to delete the partition directory, but now a create
partition comes and a new query starts to write a new delta. This should be
handled.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 557782)
Time Spent: 1h 20m (was: 1h 10m)
> Non blocking DROP PARTITION implementation
> ------------------------------------------
>
> Key: HIVE-24753
> URL: https://issues.apache.org/jira/browse/HIVE-24753
> Project: Hive
> Issue Type: New Feature
> Reporter: Zoltan Chovan
> Assignee: Zoltan Chovan
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Implement a way to execute drop partition operations in a way that doesn't
> have to wait for currently running read operations to be finished.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
