[ 
https://issues.apache.org/jira/browse/IMPALA-13491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18054283#comment-18054283
 ] 

Arnab Karmakar commented on IMPALA-13491:
-----------------------------------------

Thanks for the tip [~stigahuang]
I wanted to discuss how do we check "if the limit and fairness are enforced"?

Few thoughts come to my mind but I'd appreciate if you could choose one:
1. We can parse catalogd logs during the test and verify the sequence of permit 
acquisitions.
2. We could expose some metrics: getActiveLoadOperationsCount(), 
getLoadOperationQueueLength()
3. Query profiles analysis that contain timeline events. We can check for 
events "Acquiring load operation permit" and "Acquired load operation permit"
4. We can use some timing based tests. eg: Use debug action delay of 2 secs for 
REFRESH, trigger 6 such operations with the flag value set as 2. It should 
process the ops in 3 batches that should ideally take (3 * 2s) ~6s instead ~2s 
when all ops execute parallelly in 1 batch and assert on time.

> Add config on catalogd for controlling the number of concurrent 
> loading/refresh commands
> ----------------------------------------------------------------------------------------
>
>                 Key: IMPALA-13491
>                 URL: https://issues.apache.org/jira/browse/IMPALA-13491
>             Project: IMPALA
>          Issue Type: Improvement
>            Reporter: Manish Maheshwari
>            Assignee: Arnab Karmakar
>            Priority: Critical
>
> When running Table Loading or Refresh commands, catalogd requires working 
> memory in proportion to the number of tables been refreshed. While we have a 
> table level lock, we dont have a config to control concurrent load/refresh 
> operations.
> In case of customers that run refresh in parallel in multiple threads, the 
> number of load/refresh command can cause OOM on the catalog due to running 
> out of working memory.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to