Mailing-List: contact notifications-help@asterixdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@asterixdb.apache.org
Date: Mon, 8 May 2017 20:03:28 -0700
From: "Luo Chen (Code Review)" <do-not-reply@asterixdb.incubator.apache.org>
CC: Jenkins <jenkins@fulliautomatix.ics.uci.edu>,
	Ian Maxon <imaxon@apache.org>,
	abdullah alamoudi <bamousaa@gmail.com>,
	Jianfeng Jia <jianfeng.jia@gmail.com>,
	Yingyi Bu <buyingyi@gmail.com>, Till Westmann <tillw@apache.org>
Reply-To: cluo8@uci.edu
Subject: Change in asterixdb[master]: Implemented LSM disk components alignment
In-Reply-To: <gerrit.1493936527066.I64bf34e255def72adc73b9f87cfa628a172ea694@asterix-gerrit.ics.uci.edu>
References: <gerrit.1493936527066.I64bf34e255def72adc73b9f87cfa628a172ea694@asterix-gerrit.ics.uci.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Content-Disposition: inline
User-Agent: Gerrit/2.12.7
Message-Id: <20170509030328.A97132408CB@unhygienix.ics.uci.edu>
archived-at: Tue, 09 May 2017 03:03:33 -0000

Luo Chen has posted comments on this change.

Change subject: Implemented LSM disk components alignment
......................................................................


Patch Set 13:

(2 comments)

> (5 comments)

https://asterix-gerrit.ics.uci.edu/#/c/1725/7/hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/AbstractLSMIndex.java
File hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/AbstractLSMIndex.java:

PS7, Line 337: 
> I'm not sure that's quite right either. Why does the index need to know abo
When talking about components correlation, what it really means is that all disk components of all indexes of a dataset partition should be correlated. Indexes of different partitions of a dataset (even if they are in the same NC) could have different number of components.

Suppose a NC has two partitions P1 and P2 for a dataset. They share the same memory budget for memory components and the same PrimaryOperationTracker. Whenever P1 or P2 is full would make all indexes of P1 and P2 to be flushed. However, when memory components of P1 are full, it is possible that memory components of P2 are not flashable, e.g., they are empty or flushing, and thus there's nothing to be flushed for P2. Thus, we need to group the flush operation of all indexes of one partition, instead of all indexes of all partitions blindly.


https://asterix-gerrit.ics.uci.edu/#/c/1725/7/hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/LSMHarness.java
File hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/LSMHarness.java:

PS7, Line 458: ex
> Ok, well, maybe a comment or something to make that more explicit would be 
Done


-- 
To view, visit https://asterix-gerrit.ics.uci.edu/1725
To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I64bf34e255def72adc73b9f87cfa628a172ea694
Gerrit-PatchSet: 13
Gerrit-Project: asterixdb
Gerrit-Branch: master
Gerrit-Owner: Luo Chen <cluo8@uci.edu>
Gerrit-Reviewer: Ian Maxon <imaxon@apache.org>
Gerrit-Reviewer: Jenkins <jenkins@fulliautomatix.ics.uci.edu>
Gerrit-Reviewer: Jianfeng Jia <jianfeng.jia@gmail.com>
Gerrit-Reviewer: Luo Chen <cluo8@uci.edu>
Gerrit-Reviewer: Till Westmann <tillw@apache.org>
Gerrit-Reviewer: Yingyi Bu <buyingyi@gmail.com>
Gerrit-Reviewer: abdullah alamoudi <bamousaa@gmail.com>
Gerrit-HasComments: Yes