[jira] [Commented] (HDFS-11698) Ozone: SCM: CLI: Add Debug command

["Anu Engineer (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HDFS-11698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15993926#comment-15993926 ] 

Anu Engineer commented on HDFS-11698:
-------------------------------------

[~vagarychen] Thank you for getting this done. This undoubtedly is the most significant tool in our arsenal for ozone debugging. I am thrilled to have this tool.
+1, pending jenkins.

Few comments:
 * I like that fact that we auto detect the the DB, that is a major step in usability.
 * May be as another JIRA -- some commands to work against the sqlite DB ? Of the top of my head -- list of containers, Get Datanodes from container name
 * May be as yet another JIRA -- Another table that logs datanode UUID to container name or an index if that is possible so that we can print out all containers for a given datanode

 

> Ozone: SCM: CLI:  Add Debug command
> -----------------------------------
>
>                 Key: HDFS-11698
>                 URL: https://issues.apache.org/jira/browse/HDFS-11698
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: ozone
>    Affects Versions: HDFS-7240
>            Reporter: Anu Engineer
>            Assignee: Chen Liang
>         Attachments: HDFS-11698-HDFS-7240.001.patch
>
>
> This is not discussed as part of the standard command line set, since these are internal debug only commands.
> We have levelDB that stores key, value pairs and value is always binary blobs. These binary blobs can only be decoded using protobuf classes that are part of ozone. 
> We would like to add an extension to SCM CLI which allows us to convert levelDB to SQLlite, since SQLlite would have structured data and standard SQL will work against it. 
> The reason to use SQLlite instead of some dump file like json is the scale, it is possible that some levelDB files will have millions of keys, so having an SQL DB with proper index will make debugging much easier.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

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