[
https://issues.apache.org/jira/browse/HBASE-17333?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15760339#comment-15760339
]
ramkrishna.s.vasudevan commented on HBASE-17333:
------------------------------------------------
bq.The issue is this. When read from conf, the default to be used as
HColumnDescriptor.MemoryCompaction.NONE
That single change should fix it.
I don't want to change that. Let the default be BASIC only. If there is no HCD
setting then we should ensure we go with default memstore. Since
{code}
public HColumnDescriptor setInMemoryCompaction(MemoryCompaction
inMemoryCompaction) {
{code}
accepts a MemoryCompaciton who ever is using it will be setting
NONE/BASIC/EAGER. There there is no problem. Only when some one is not using it
explicitly we get this problem. So I think the current way in patch is fine.
bq.Or may be after this commit, that can be closed and we can open a new one to
track, why the test fail with CompactingMS feature
Ya we can do that. Raise another JIRA and dig there.
> HBASE-17294 always ensures CompactingMemstore is default
> --------------------------------------------------------
>
> Key: HBASE-17333
> URL: https://issues.apache.org/jira/browse/HBASE-17333
> Project: HBase
> Issue Type: Bug
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Priority: Critical
> Fix For: 2.0.0
>
> Attachments: HBASE-17333.patch
>
>
> Was the purpose of HBASE-17294 is to make Compacting Memstore as default? Am
> not sure on that. But that patch makes DefaultMemstore as a Noop. This JIRA
> is to discuss and revert back to default memstore only if the family is not
> configured for in memory flush/compaction.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
