phoenix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ohad Shacham (JIRA)" <>
Subject [jira] [Commented] (PHOENIX-4278) Implement pure client side transactional index maintenance
Date Thu, 01 Feb 2018 08:02:01 GMT


Ohad Shacham commented on PHOENIX-4278:

[~jamestaylor], As per our discussion I applied the patch on-top-of of 4.x-HBase-1.3 and
ran testing. It terminated with the same errors I received when running 4.x-HBase-1.3 without
my patch. 

This are the errors:

[*ERROR*] *Failures:* 

[*ERROR*]   *AggregateIT.testAvgGroupByOrderPreservingWithStats:432 expected:<13> but


[*ERROR*] *Tests run: 3365, Failures: 1, Errors: 0, Skipped: 7*


[*ERROR*] *Failures:* 

[*ERROR*] *org.apache.phoenix.end2end.index.MutableIndexFailureIT.testIndexWriteFailure[MutableIndexFailureIT_transactional=false,localIndex=false,isNamespaceMapped=false,disableIndexOnWriteFailure=true,failRebuildTask=false,throwIndexWriteFailure=null](org.apache.phoenix.end2end.index.MutableIndexFailureIT)*

[*ERROR*]   *Run 1: MutableIndexFailureIT.testIndexWriteFailure:345->checkStateAfterRebuild:389*

[*INFO*]   *Run 2: PASS*

[*INFO*]   *Run 3: PASS*


[*ERROR*]   *PartialIndexRebuilderIT.testConcurrentUpsertsWithRebuild:221 Ran out of time*

[*ERROR*]   *PartialIndexRebuilderIT.testMultiVersionsAfterFailure:461*

[*ERROR*]   *PartialIndexRebuilderIT.testUpsertNullTwiceAfterFailure:521*


[*ERROR*] *Tests run: 600, Failures: 4, Errors: 0, Skipped: 44*


> Implement pure client side transactional index maintenance
> ----------------------------------------------------------
>                 Key: PHOENIX-4278
>                 URL:
>             Project: Phoenix
>          Issue Type: Improvement
>            Reporter: James Taylor
>            Assignee: Ohad Shacham
>            Priority: Major
> The index maintenance for transactions follows the same model as non transactional tables
- coprocessor based on data table updates that looks up previous row value to perform maintenance.
This is necessary for non transactional tables to ensure the rows are locked so that a consistent
view may be obtained. However, for transactional tables, the time stamp oracle ensures uniqueness
of time stamps (via transaction IDs) and the filtering handles a scan seeing the "true" last
committed value for a row. Thus, there's no hard dependency to perform this on the server
> Moving the index maintenance to the client side would prevent any RS->RS RPC calls
(which have proved to be troublesome for HBase). It would require returning more data to the
client (i.e. the prior row value), but this seems like a reasonable tradeoff.

This message was sent by Atlassian JIRA

View raw message