hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anu Engineer (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-11582) Block Storage : add SCSI target access daemon
Date Mon, 27 Mar 2017 22:58:42 GMT

    [ https://issues.apache.org/jira/browse/HDFS-11582?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15944192#comment-15944192

Anu Engineer commented on HDFS-11582:

[~vagarychen] Thanks for adding this. This makes CBlock starting to look real. Some minor

* File : {{CBlockClientServerProtocolClientSideTranslatorPB.java}}
I am struggling with this name. It has lots of words but no meaning really pops up in my mind.
Can we please rename it ? 

* Class comment: 
{{* The client side of CBlockClientServerProtcol}}, may be add some more information here
about what this class really does.

* Function: {{mountVolume}}
-- Should we check if userName and volumeName are not null ? 

* Should we have an else part that logs error and throws an exception ?  
{{if (containerID.hasPipeline()) // it should always have a pipeline only except for tests.}}

* Just a curiosity question. You can ignore this since it is very hypothetical.
 {{if (!targets.containsKey(volumeKey))}} , we have this code in {{isValidTargetName}}
 who ensures that we don't have race condition that is two volumeNames from 2 different
 CBlock servers are not racing here ? I understand that it is a pretty far fetched scenerio
and I was wondering if the jscsi target server does any kind of concurrency control.

*  File: {{SCSITargetDaemon.java}}
 ContainerOperationClient.setContainerSizeB(containerSizeGB * 1024 * 1024 * 1024L);
 replace with 
 ContainerOperationClient.setContainerSizeB(containerSizeGB * OzoneConsts.GB);

* Debug Statements ? 
These look like statements that came out of debugging. Could you please remove it ? 
ozoneConf.setBoolean(OzoneConfigKeys.OZONE_ENABLED, true);
ozoneConf.setBoolean(OzoneConfigKeys.OZONE_TRACE_ENABLED_KEY, true);
ozoneConf.set(OzoneConfigKeys.OZONE_HANDLER_TYPE_KEY, "distributed")
I understand that these are required for CBlock to work, but shouldn't we just assert or log
error and get out instead of setting these values ourselves.

> Block Storage : add SCSI target access daemon
> ---------------------------------------------
>                 Key: HDFS-11582
>                 URL: https://issues.apache.org/jira/browse/HDFS-11582
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: hdfs
>            Reporter: Chen Liang
>            Assignee: Chen Liang
>         Attachments: HDFS-11582-HDFS-7240.001.patch, HDFS-11582-HDFS-7240.002.patch
> This JIRA adds the daemon process that exposes SCSI target access. More specifically,
with this daemon process running, any OS with SCSI can talk to this daemon process and treat
CBlock volumes as SCSI targets, in this way the user can mount the volume just like the POSIX

This message was sent by Atlassian JIRA

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

View raw message