[ 
https://issues.apache.org/jira/browse/HBASE-532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12588880#action_12588880
 ] 

stack commented on HBASE-532:
-----------------------------

On first comment, its snapshot is going to be called again because flush is 
going to fail; the way things are currently, we log warning that we're being 
recalled but we go again to try and flush what was in snapshot.  Your 
suggestion would make we'd throw an exception and then this region would dead 
in the water, at least as far as flushing is concerned.

On clearSnapshot, yeah, just a precaution.  Unlikely but if it happened, we'd 
want it logged.

Adding the 'this' is minor potatoes.  Yes, its stylistic.  I like them -- easy 
way telling data members from locals -- but I can see you have been actively 
removing them.

Onk getNextRow not suppressing deletes, yes it should. I was thinking we'd let 
the result out and then they'd be merged higher up but thats wrong behavior 
here -- scope is memcache scanner only.

On Long.valueOf, gives chance of internal Long cache returning a ready-made 
long -- if it exists (also suppresses an eclipse warning to the effect that the 
long is being auto-boxed up into a Long object).

Making a fix for getNextRow now...

> Odd interaction between HRegion.get, HRegion.deleteAll and compactions
> ----------------------------------------------------------------------
>
>                 Key: HBASE-532
>                 URL: https://issues.apache.org/jira/browse/HBASE-532
>             Project: Hadoop HBase
>          Issue Type: Bug
>    Affects Versions: 0.2.0, 0.1.1
>            Reporter: Jim Kellerman
>            Assignee: stack
>            Priority: Blocker
>             Fix For: 0.2.0, 0.1.2
>
>         Attachments: 532.patch
>
>
> If you apply the patch for HBASE-483 to the 0.1 branch and comment out lines 
> 309 and 315 of MetaUtils.java (which force compactions of the root and meta 
> regions respectively), TestMergeTool fails. Why forcing compactions makes the 
> test succeed is a mystery to me.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to