[
https://issues.apache.org/jira/browse/HBASE-14918?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15121385#comment-15121385
]
Eshcar Hillel commented on HBASE-14918:
---------------------------------------
I reviewed the mslab-move patch. Software-engineering-wise I am not at all
convinced that the right place for mslab is in HStore level.
The compacting memstore is an example in which cells are allocated at the
memstore level and not the store level.
But more important is what you say about off-heap memory. I have no experience
with off-heaping.
Can you please elaborate why the suggested design cannot be off-heap, and what
is needed to allow it be off-heap?
In addition, you refer to the write-path, but actually the write-path goes
through mutable-segment that stores the data in a CSLM format.
Only reads and scans access the cell block.
It is good we have this discussion at this point since it relates to the design
of task #4, and can also affect task #3.
However, [~stack], is there anything that prevents committing the patch of task
#1. Is it not committed due to the MSLAB issue?
IMO, the mslab is orthogonal to task #1. If it is decided that it needs to
move, then it is possible to do so even after the patch.
> In-Memory MemStore Flush and Compaction
> ---------------------------------------
>
> Key: HBASE-14918
> URL: https://issues.apache.org/jira/browse/HBASE-14918
> Project: HBase
> Issue Type: Umbrella
> Affects Versions: 2.0.0
> Reporter: Eshcar Hillel
> Assignee: Eshcar Hillel
> Fix For: 0.98.18
>
> Attachments: CellBlocksSegmentDesign.pdf, MSLABMove.patch
>
>
> A memstore serves as the in-memory component of a store unit, absorbing all
> updates to the store. From time to time these updates are flushed to a file
> on disk, where they are compacted (by eliminating redundancies) and
> compressed (i.e., written in a compressed format to reduce their storage
> size).
> We aim to speed up data access, and therefore suggest to apply in-memory
> memstore flush. That is to flush the active in-memory segment into an
> intermediate buffer where it can be accessed by the application. Data in the
> buffer is subject to compaction and can be stored in any format that allows
> it to take up smaller space in RAM. The less space the buffer consumes the
> longer it can reside in memory before data is flushed to disk, resulting in
> better performance.
> Specifically, the optimization is beneficial for workloads with
> medium-to-high key churn which incur many redundant cells, like persistent
> messaging.
> We suggest to structure the solution as 4 subtasks (respectively, patches).
> (1) Infrastructure - refactoring of the MemStore hierarchy, introducing
> segment (StoreSegment) as first-class citizen, and decoupling memstore
> scanner from the memstore implementation;
> (2) Adding StoreServices facility at the region level to allow memstores
> update region counters and access region level synchronization mechanism;
> (3) Implementation of a new memstore (CompactingMemstore) with non-optimized
> immutable segment representation, and
> (4) Memory optimization including compressed format representation and off
> heap allocations.
> This Jira continues the discussion in HBASE-13408.
> Design documents, evaluation results and previous patches can be found in
> HBASE-13408.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
