hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kurtis Heimerl (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-1298) adding user info to file
Date Thu, 03 May 2007 22:49:15 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-1298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12493547

Kurtis Heimerl commented on HADOOP-1298:

I like point 1. I'll change that once we figure out point two. it might be meaningful to deprecate
the existing length, is directory and stuff. 

Also, I reverted change 7 from the previous list. I've decided retuning null makes more sense.
File permissions at this layer are entirely client side. The server contains no file permissions
so returning some would be faulty. The permissions at this layer are not going to be used
to for access control anyhow. 

As far as point two, I agree that the existing design isn't right. It's very error prone,
moreso if I add more fields as per 1. 

These solutions bother me as being too closely tied to DFS. The fact that the client and server
contain almost identical data structures indicates too much coupling. It would be difficult
to implement another permissions model with this as the front end. They'd have to translate
everything to these ints. 

In this layer these simply are fields. I think it's reasonable to pack the booleans into an
int, but it should be seperated from the file system. No one will ever instantiate this class
as you two have listed. It's going to be:

FS_Specific_Permissions perm = fs.getPermissions(f);
FileStatus s = new FileStatus(perm.something(), perm.somethingElse())

There's got to be a better design to this. Do you think it would be reasonable to have two
constructors, one taking the booleans packed into an int and one taking the booleans? I don't
want the int to mean anything more than the booleans. 

Lastly, someone should look at the groups code. I don't like it and if someone came up with
something better, I think that'd be great. 

> adding user info to file
> ------------------------
>                 Key: HADOOP-1298
>                 URL: https://issues.apache.org/jira/browse/HADOOP-1298
>             Project: Hadoop
>          Issue Type: New Feature
>          Components: dfs, fs
>            Reporter: Kurtis Heimerl
>         Attachments: hadoop-user-munncha.patch, hadoop-user-munncha.patch, hadoop-user-munncha.patch,
hadoop-user-munncha.patch4, hadoop-user-munncha.patch5, hadoop-user-munncha.patch6, hadoop-user-munncha.patch7
> I'm working on adding a permissions model to hadoop's DFS. The first step is this change,
which associates user info with files. Following this I'll assoicate permissions info, then
block methods based on that user info, then authorization of the user info. 
> So, right now i've implemented adding user info to files. I'm looking for feedback before
I clean this up and make it offical. 
> I wasn't sure what release, i'm working off trunk. 

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message