[
https://issues.apache.org/jira/browse/HBASE-15082?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HBASE-15082:
--------------------------
Attachment: 15082v2.patch
Findbugs complaints are not from this patch. Trying to address in another
issue. The rest is mine though. Here is another attempt.
{code}
if a thread A begins a transaction writing row A and then thread B beings a
transaction on row B, but then thread B completes and thread A does not...
A subsequent read of row B will not see the most recent write on this row
because the read point has not caught up because the outstanding mvcc write
from thread A will hold up the complete of the thread B mvcc complete; its
this latter that will catch up the mvcc read point. In branch-1, we did
a read uncommitted to get around this fact. TODO: Add test here.
This patch is a forward port of the branch-1 'fix'. Also does refactor
of a few methods that were copy/pastes of each other: increment and append
as well as checkAndMutate and checkAndRowMutation.
M b/hbase-client/src/main/java/org/apache/hadoop/hbase/client/Scan.java
Pass on any isolation level specified.
M b/hbase-client/src/test/java/org/apache/hadoop/hbase/client/TestIncrement.java
Give test a proper name.
M b/hbase-common/src/main/java/org/apache/hadoop/hbase/TagUtil.java
Added utility methods carryForwardTags and carryForwardTTLTag.
M
b/hbase-examples/src/main/java/org/apache/hadoop/hbase/coprocessor/example/ZooKeeperScanPolicyObserver.java
Fix method capitalization
M b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.java
Move all mvcc completes and syncs under row lock. While I'm at it, since sync
is
now under row lock, move update of memstore AFTER sync-to-WAL. Allows for
simplification
Removed guts of a DLR method, setRecovering. Doesn't work and unsupported
since 1.2.
Fixed capitialization on getReadpoint.
Tried to break up a few massive methods.
D
b/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/wal/ReplayHLogKey.java
Removed. From DLR.
M b/hbase-server/src/main/java/org/apache/hadoop/hbase/wal/WALKey.java
Add utility completeMvcc
A
b/hbase-server/src/test/java/org/apache/hadoop/hbase/IncrementPerformanceTest.java
Added simple increment test.
{code}
> Fix merge of MVCC and SequenceID performance regression
> -------------------------------------------------------
>
> Key: HBASE-15082
> URL: https://issues.apache.org/jira/browse/HBASE-15082
> Project: HBase
> Issue Type: Sub-task
> Components: Performance
> Reporter: stack
> Assignee: stack
> Priority: Critical
> Fix For: 2.0.0, 1.2.0, 1.3.0
>
> Attachments: 15082.patch, 15082v2.patch, 15082v2.txt, 15082v3.txt,
> 15082v4.patch, 15082v5.patch
>
>
> This is general fix for increments (appends, checkAnd*) perf-regression
> identified in the parent issue. HBASE-15031 has a narrow fix for branch-1.1
> and branch-1.0.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
