phoenix-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Purtell (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (PHOENIX-4076) Move master branch up to HBase 1.4.0-SNAPSHOT
Date Mon, 23 Oct 2017 19:33:00 GMT

    [ https://issues.apache.org/jira/browse/PHOENIX-4076?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16215715#comment-16215715
] 

Andrew Purtell edited comment on PHOENIX-4076 at 10/23/17 7:32 PM:
-------------------------------------------------------------------

I ran the test suite against latest 1.3.2-SNAPSHOT, and applied this patch and ran it again
against latest 1.4.0-SNAPSHOT.

The 1.3.2-SNAPSHOT results:
{noformat}
HBaseManagedTimeTests

[ERROR] Failures: 
[ERROR]   SequenceIT.testDuplicateSequences:102 Duplicate sequences
[ERROR] Errors: 
[ERROR]   SequenceIT.testSequenceDefault:761 » SequenceAlreadyExists ERROR 1200 (42Z00):...

ParallelStatsEnabledTests

[ERROR] Failures: 
[ERROR]   PartialIndexRebuilderIT.testConcurrentUpsertsWithRebuild:220 Ran out of time
{noformat}

The 1.4.0-SNAPSHOT results:
{noformat}
HBaseManagedTimeTests

[ERROR] Failures: 
[ERROR]   SequenceIT.testDuplicateSequences:102 Duplicate sequences
[ERROR] Errors: 
[ERROR]   SequenceIT.testSequenceDefault:761 » SequenceAlreadyExists ERROR 1200 (42Z00):...

ParallelStatsEnabledTest

[ERROR] Failures: 
[ERROR]   LocalIndexSplitMergeIT.testLocalIndexScanAfterRegionsMerge:236 expected:<[h]>
but was:<[i]>
{noformat}

The results are, basically, identical in terms of number of failed tests. Using 1.4.0-SNAPSHOT
instead of 1.3.2-SNAPSHOT does not appear to introduce more instability. The local index test
failure on 1.4.0-SNAPSHOT may be germane to changes impacting that feature specifically.


was (Author: apurtell):
I ran the test suite against latest 1.3.2-SNAPSHOT, and applied this patch and ran it again
against latest 1.4.0-SNAPSHOT.

The 1.3.2-SNAPSHOT results:
{noformat}
HBaseManagedTimeTests

[ERROR] Failures: 
[ERROR]   SequenceIT.testDuplicateSequences:102 Duplicate sequences
[ERROR] Errors: 
[ERROR]   SequenceIT.testSequenceDefault:761 » SequenceAlreadyExists ERROR 1200 (42Z00):...

ParallelStatsEnabledTests

[ERROR] Failures: 
[ERROR]   PartialIndexRebuilderIT.testConcurrentUpsertsWithRebuild:220 Ran out of time
{noformat}

The 1.4.0-SNAPSHOT results:
{noformat}
HBaseManagedTimeTests

[ERROR] Failures: 
[ERROR]   SequenceIT.testDuplicateSequences:102 Duplicate sequences
[ERROR] Errors: 
[ERROR]   SequenceIT.testSequenceDefault:761 » SequenceAlreadyExists ERROR 1200 (42Z00):...

ParallelStatsEnabledTest

[ERROR] Failures: 
[ERROR]   LocalIndexSplitMergeIT.testLocalIndexScanAfterRegionsMerge:236 expected:<[h]>
but was:<[i]>
{noformat}

The results are, basically, identical. 

> Move master branch up to HBase 1.4.0-SNAPSHOT
> ---------------------------------------------
>
>                 Key: PHOENIX-4076
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-4076
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: Andrew Purtell
>            Assignee: Andrew Purtell
>         Attachments: PHOENIX-4076.patch
>
>
> Move master branch up to HBase 1.4.0-SNAPSHOT. 
> There are some compilation problems. 
> Valid compatibility breaks are addressed and fixed by HBASE-18431. This analysis is a
compilation attempt of Phoenix master branch against 1.4.0-SNAPSHOT artifacts including the
HBASE-18431 changes. 
> HBASE-16584 removed PayloadCarryingRpcController, breaking compilation of MetadataRpcController,
InterRegionServerIndexRpcControllerFactory, IndexRpcController, ClientRpcControllerFactory,
and InterRegionServerMetadataRpcControllerFactory. This class was annotated as Private so
was fair game to remove. It will be gone in HBase 1.4.x and up. 
> DelegateRegionObserver needs to implement added interface method postCommitStoreFile.
> DelegateHTable, TephraTransactionTable, and OmidTransactionTable need to implement added
interface methods for getting and setting read and write timeouts. 
> PhoenixRpcScheduler needs to implement added interface methods for getting handler counts.

> Store file readers/writers/scanners have been refactored and the local index implementation,
which implements or overrides parts of this refactored hierarchy will have to also be refactored.
> DelegateRegionCoprocessorEnvironment needs to implement added method getMetricRegistryForRegionServer
> Another issue with IndexRpcController: incompatible types: int cannot be converted to
org.apache.hadoop.hbase.TableName



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message