[
https://issues.apache.org/jira/browse/HBASE-16438?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15943108#comment-15943108
]
Anastasia Braginsky commented on HBASE-16438:
---------------------------------------------
bq. But am not sure how you mean this hardening it again.
public class SoftReference<T> extends Reference<T>
So I think you can use just reference in the map type and later just to update
the map with new mapping (delete old mapping from some chunkID to soft
reference and add a new mapping from same chunkID to hard reference). If this
doesn't work for some reason we may have two maps <Long, SoftRef(Chunk)> and
<Long, HardRef(Chunk)>.
bq. Now when a segment is getting converted from CSLM based into CellChunkMap
based, can we just track the chunkIds actually getting used? (Which cells
getting moved into ChunkMap). All the remaining chunks seems of no use and can
be immediate removed from the map.
Nice idea. I think the result of this idea is the same as combining soft and
hard pointers in the Creator's Map. However, with combining soft and hard
pointers in the Creator's Map the chunks can be GCed also in active segment.
bq. Do you feel that can work? I have implemented the same in a patch but not
posted it as I need to see if there is really a perf penalty.
I saw this idea, I feel it can work, I think it is a bit complicated and may
cost us some performance. I believe we can achieve the same for less. But if
you want to go for it, I am OK with that.
> Create a cell type so that chunk id is embedded in it
> -----------------------------------------------------
>
> Key: HBASE-16438
> URL: https://issues.apache.org/jira/browse/HBASE-16438
> Project: HBase
> Issue Type: Sub-task
> Affects Versions: 2.0.0
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Attachments: HBASE-16438_1.patch,
> HBASE-16438_3_ChunkCreatorwrappingChunkPool.patch,
> HBASE-16438_4_ChunkCreatorwrappingChunkPool.patch, HBASE-16438.patch,
> MemstoreChunkCell_memstoreChunkCreator_oldversion.patch,
> MemstoreChunkCell_trunk.patch
>
>
> For CellChunkMap we may need a cell such that the chunk out of which it was
> created, the id of the chunk be embedded in it so that when doing flattening
> we can use the chunk id as a meta data. More details will follow once the
> initial tasks are completed.
> Why we need to embed the chunkid in the Cell is described by [~anastas] in
> this remark over in parent issue
> https://issues.apache.org/jira/browse/HBASE-14921?focusedCommentId=15244119&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15244119
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
