You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Anoop Sam John (JIRA)" <ji...@apache.org> on 2016/07/11 07:50:11 UTC

[jira] [Comment Edited] (HBASE-16195) Should not add chunk into chunkQueue if not using chunk pool in HeapMemStoreLAB

    [ https://issues.apache.org/jira/browse/HBASE-16195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15370333#comment-15370333 ] 

Anoop Sam John edited comment on HBASE-16195 at 7/11/16 7:49 AM:
-----------------------------------------------------------------

{code}
 long memstoreFlushSize = conf.getLong(HConstants.HREGION_MEMSTORE_FLUSH_SIZE,
101	        HTableDescriptor.DEFAULT_MEMSTORE_FLUSH_SIZE);
102	      // by default max heap size of chunk queue won't exceed memstoreFlushSize
103	      int queueLength = conf.getInt(CHUNK_QUEUE_LENGTH, (int) (memstoreFlushSize / chunkSize));
104	      chunkQueue = new LinkedBlockingQueue<Chunk>(queueLength);
{code}
Why we need this change also? I mean sizing the Q.  When pool is not in place, we avoid adding to this Q . That should be enough?
This calculation of the size based on the flush size can get wrong.  The flush size is not a hard limit. We allow the memstore to go bigger by 4 times over the default value.  (default flush size is 128MB).   So any sort of calc we do here can go wrong!  


was (Author: anoop.hbase):
{code}
 long memstoreFlushSize = conf.getLong(HConstants.HREGION_MEMSTORE_FLUSH_SIZE,
101	        HTableDescriptor.DEFAULT_MEMSTORE_FLUSH_SIZE);
102	      // by default max heap size of chunk queue won't exceed memstoreFlushSize
103	      int queueLength = conf.getInt(CHUNK_QUEUE_LENGTH, (int) (memstoreFlushSize / chunkSize));
104	      chunkQueue = new LinkedBlockingQueue<Chunk>(queueLength);
{code}
Why we need this change also? I mean sizing the Q.  When pool is not in place, we avoid adding to this Q . That should be enough?
This calculation of the size based on the flush size can get wrong.  The flush size is not a hard limit. We allow the memstore to go bigger by 4 times by default.  (default flush size is 128MB).   So any sort of calc we do here can go wrong!  

> Should not add chunk into chunkQueue if not using chunk pool in HeapMemStoreLAB
> -------------------------------------------------------------------------------
>
>                 Key: HBASE-16195
>                 URL: https://issues.apache.org/jira/browse/HBASE-16195
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Yu Li
>            Assignee: Yu Li
>         Attachments: HBASE-16195.patch, HBASE-16195_v2.patch, HBASE-16195_v3.patch
>
>
> Problem description and analysis please refer to HBASE-16193



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)