hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mukul Kumar Singh (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-12199) Ozone: OzoneFileSystem: OzoneFileystem initialization code
Date Fri, 11 Aug 2017 11:26:00 GMT

    [ https://issues.apache.org/jira/browse/HDFS-12199?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16123213#comment-16123213
] 

Mukul Kumar Singh commented on HDFS-12199:
------------------------------------------

Thanks for the review [~xyao],

TestOzoneFileInterfaces.java
Line 57/89: we need to ensure storageHandler is closed properly after the test.
.bq done

hadoop-tools/hadoop-ozone/pom.xml
Line 110-112: do we need dependencies for RocksDb as well?
bq. I have removed the leveldb dependency as well, as there is no requirement for leveldb
here.
If it will be needed, we can add the dependency later.

OzoneFileSystem.java
Line 156-204: empty implementation. I assume this will be done in followup JIRAs.
.bq, yes the implementation will followup in the next jiras

Thanks for the review [~cheersyang], Please find my replies inline.

Can we name the schema to oz instead of ozfs ? I checked some existing FS implementations,
their schema are like, "oss", "adl", "s3a" etc. none of them contains "fs" suffix.
bq. Other filesystems in hdfs-project are named "hdfs, viewfs, webhdfs" etc and hence I chose
the name :). Also in the hdfs cli, we have an "oz" command which is meant only for ozone operation.
I wanted to name this scheme fs to distinguish this as a filesystem URI.

1. Since oz FS is initiated per bucket, we need to ensure bucket existence during the initialize.
But there seems no code to check this.
bq. This check will be done as part of the following lines 107-108, which will execute a HTTP
query
{code}
      OzoneVolume volume = ozone.getVolume(volumeStr);
      this.bucket = volume.getBucket(bucketStr);
{code}

2. line 84 - 88, question: if a client passes, e.g /vol/bucket/ppp, as the URI path to init
oz FS, is this a valid configuration or an exception to throw?
bq. That is a valid URI, the volume and bucket will be extracted out of this URI. This will
also be needed to support fully qualified ozone fs URI

3. line 99, line 109: can we move constants "http://" and "/user" to Constants class?
bq. done

4. Can you implement at least one FS API in OzoneFileSystem to prove this is working?
bq. I have implemented all the basic APIs(list, read, write, mkdir, getfilestatus) and they
are working, I will post those in the subsequent patches.



> Ozone: OzoneFileSystem: OzoneFileystem initialization code
> ----------------------------------------------------------
>
>                 Key: HDFS-12199
>                 URL: https://issues.apache.org/jira/browse/HDFS-12199
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: ozone
>    Affects Versions: HDFS-7240
>            Reporter: Mukul Kumar Singh
>            Assignee: Mukul Kumar Singh
>             Fix For: HDFS-7240
>
>         Attachments: HDFS-12199-HDFS-7240.001.patch
>
>
> This jira will be used to add OzoneFileySystem initialization code. This jira is based
out of HDFS-11704. I will attach a design document to HDFS-11704.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org


Mime
View raw message