Return-Path: X-Original-To: apmail-hadoop-hdfs-issues-archive@minotaur.apache.org Delivered-To: apmail-hadoop-hdfs-issues-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id AFFC317728 for ; Thu, 18 Jun 2015 08:01:01 +0000 (UTC) Received: (qmail 35286 invoked by uid 500); 18 Jun 2015 08:01:00 -0000 Delivered-To: apmail-hadoop-hdfs-issues-archive@hadoop.apache.org Received: (qmail 35232 invoked by uid 500); 18 Jun 2015 08:01:00 -0000 Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: hdfs-issues@hadoop.apache.org Delivered-To: mailing list hdfs-issues@hadoop.apache.org Received: (qmail 35220 invoked by uid 99); 18 Jun 2015 08:01:00 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 18 Jun 2015 08:01:00 +0000 Date: Thu, 18 Jun 2015 08:01:00 +0000 (UTC) From: "Jitendra Nath Pandey (JIRA)" To: hdfs-issues@hadoop.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (HDFS-8457) Ozone: Refactor FsDatasetSpi to pull up HDFS-agnostic functionality into parent interface MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HDFS-8457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14591433#comment-14591433 ] Jitendra Nath Pandey commented on HDFS-8457: -------------------------------------------- [~arpitagarwal], this looks like a great patch. A few questions and comments: # {{private boolean datasetIsFsDataset = false;}} : We could just check 'instanceof' instead of having this variable, is there a performance concern here? Did you consider delegating cacheReport or some parts of it to dataset implementation, that would reduce the need for this boolean check? # I think getBlockLocalPathInfo assumes that it must be FsDatasetSpi, which is correct. please add a comment for that. The same is true for a few more methods. # What is the reason for not moving {{public List getFinalizedBlocks(String bpid)}} to DatasetSpi? # {{public boolean contains(ExtendedBlock block)}} can possibly move to DataSetSpi. # {{FsVolumeSpi}} and {{VolumeSpi}} : getPath is moved to VolumeSpi while getBasePath continues to be in FsVolumeSpi, is that intended? # {{public long getAvailable() throws IOException}} should also move to VolumeSpi. # DataNodeTestUtils.java has only changes to imports. > Ozone: Refactor FsDatasetSpi to pull up HDFS-agnostic functionality into parent interface > ----------------------------------------------------------------------------------------- > > Key: HDFS-8457 > URL: https://issues.apache.org/jira/browse/HDFS-8457 > Project: Hadoop HDFS > Issue Type: Sub-task > Components: datanode > Reporter: Arpit Agarwal > Assignee: Arpit Agarwal > Attachments: HDFS-8457-HDFS-7240.01.patch, HDFS-8457-HDFS-7240.02.patch, HDFS-8457-HDFS-7240.03.patch, HDFS-8457-HDFS-7240.04.patch, HDFS-8457-HDFS-7240.05.patch, HDFS-8457-HDFS-7240.06.patch > > > FsDatasetSpi can be split up into HDFS-specific and HDFS-agnostic parts. The HDFS-specific parts can continue to be retained in FsDataSpi while those relating to volume management, block pools and upgrade can be moved to a parent interface. > There will be no change to implementations of FsDatasetSpi. -- This message was sent by Atlassian JIRA (v6.3.4#6332)