[
https://issues.apache.org/jira/browse/HBASE-16438?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15907440#comment-15907440
]
Anastasia Braginsky commented on HBASE-16438:
---------------------------------------------
Hey [~ram_krish]!
Thanks for publishing the new patch! I have looked on it, but didn't publish my
detailed comments yet. Here are some of my general thoughts:
1. I agree with [~ram_krish] that we should exclude CellChunkMap implementation
from this patch. This patch is big enough and we have here enough design and
implementation details. So we should better concentrate on making just this in
a great way. My only concern is that if we remove the CellChunkMap we have less
abilities to test the new code. We should make sure the new code is well
tested. Please add the new tests for all the relevant interfaces (getChunkId()
from any cell, translate chunk ID into chunk, etc.).
2. I disagree with the design decision of putting MemStoreLABChunkCreator to be
part of MemStoreChunkPool. I also don't understand why MemStoreLABChunkCreator
works only with chunks allocated from pool. Please decouple the ChunkPool and
the ChunkCreator, there can be chunks without pool, so there should exist
ChunkCreator even if it was decided not to allocate/de-allocate Chunks via
Pool. Whether it is reasonable or not to have chunks without pool is a separate
question and a matter of usage. As we are not sure how exactly CellChunkMap is
going to be used in the future we should not block any option for now. As I
have already said, the ChunkCreator should be responsible for ANY chunk
allocation and ANY chunk should have its ID. Maybe it makes sense to do some
small design document?
Now I am going to publish more detailed comments on the RB itself.
[~ram_krish], thanks once again!
> 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: 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)
