asterixdb-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ildar Absalyamov (Code Review)" <>
Subject Change in asterixdb[master]: [NO ISSUE][STO][IDX] LSM storage cleanup
Date Wed, 27 Sep 2017 23:42:12 GMT
Ildar Absalyamov has posted comments on this change.

Change subject: [NO ISSUE][STO][IDX] LSM storage cleanup

Patch Set 17:

(1 comment)
File hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/

PS17, Line 26: IChainedComponentBulkLoader
> AsterixDB Tuple Compactor works in AsterixDB level.
I see why you would like to have such factory. Let's pretend that we have one. You will still
need to pass it somehow to LSMBTreeDiskComponent somehow, right? That means another argument
in LSMBTreeDiskComponentFactory, which is instantiated in one of LSMBTreeUtil.createLSMTree()
static method (an entry point of LSMBTree instantiating). Feels like introducing another parallel
set of factory classes is too much, we already have one for DiskComponents and their DiskComponentsFactories
Instead, I suggest that you extend LSMBTreeDiskComponent to create a new class (maybe only
overriding createComponentBulkLoader() method with your logic) and appropriate *Factory and
use it in LSMBTreeUtil.createLSMTree() where you will do initialization anyways. I believe
this is much better than passing Factory objects through multiple calls.

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I6739d751b990e7a28e03e32a5de6e2b670d37a1e
Gerrit-PatchSet: 17
Gerrit-Project: asterixdb
Gerrit-Branch: master
Gerrit-Owner: Ildar Absalyamov <>
Gerrit-Reviewer: Anon. E. Moose #1000171
Gerrit-Reviewer: Ian Maxon <>
Gerrit-Reviewer: Ildar Absalyamov <>
Gerrit-Reviewer: Jenkins <>
Gerrit-Reviewer: Luo Chen <>
Gerrit-Reviewer: Till Westmann <>
Gerrit-Reviewer: Wail Alkowaileet <>
Gerrit-Reviewer: abdullah alamoudi <>
Gerrit-HasComments: Yes

View raw message