[Gluster-users] help, chmod g+s FOLDER doesn't make sense under glusterfs

[Dai, Manhong]

Hi Raghavendra,


	I appreciate your quick reply.


	I would hope that glusterfs can set the file's group id to the mode of the
parent directory(if it has a set group id). This feature is very useful for group sharing. Let me show an example here.


[citiceo at bass1 tmp]$ id
uid=10000(citiceo) gid=1000(citi) groups=1000(citi),1002(allbankusers)
[citiceo at bass1 tmp]$ ll -d a
drwxrws--- 2 root allbankusers 4096 Aug 25 22:41 a
[citiceo at bass1 tmp]$ echo abc > a/b
[citiceo at bass1 tmp]$ ll a/b
-rw-r----- 1 citiceo allbankusers 4 Aug 25 22:41 a/b

[chaseceo at bass1 tmp]$ id
uid=10001(chaseceo) gid=1003(chase) groups=1002(allbankusers),1003(chase)
[chaseceo at bass1 tmp]$ cat a/b
abc


	This examples shows how citiceo share a file to chaseceo under ext3 filesystem. If glusterfs is used, citiceo has to run an additional command chgrp to change the file's group ownership.


	chgrp is not a hassle for any cluster administrator. But for regular users, it is a big problem. Moreover, glusterfs is a great filesystem, with which one can create a huge size file system for a cluster that has complicated users and groups mapping. Group sharing is an important feature for such a system, so cluster administrators don't need to train their users.


Best,
Manhong



-----Original Message-----
From: Raghavendra G [mailto:raghavendra.hg at gmail.com]
Sent: Mon 8/25/2008 10:13 PM
To: Dai, Manhong
Cc: gluster-users at gluster.org
Subject: Re: [Gluster-users] help, chmod g+s FOLDER doesn't make sense under glusterfs
 
Hi,

below is an excerpt of 'man creat'

O_CREAT
              If the file does not exist it will be created.  The owner
(user ID) of the file is set to
              the effective user ID of the process.  *The group ownership
(group ID) is  set  either  to
              the effective group ID of the process or to the group ID of
the parent directory* (depend-
              ing on filesystem type and mount options, and the mode of the
parent directory, see,  for
              example,  the mount options bsdgroups and sysvgroups of the
ext2 filesystem, as described
              in mount(8)).

glusterfs is setting the group ownership to effective group id of the
process.

regards,

On Tue, Aug 26, 2008 at 5:51 AM, Dai, Manhong <daimh at umich.edu> wrote:

>  Hi,
>
>
> This test is done under ext3 file system
> [root at bass1 ~]# id
> uid=0(root) gid=0(root)
> groups=0(root),1(bin),2(daemon),3(sys),4(adm),6(disk),10(wheel)
> [root at bass1 ~]# ll -d a
> drwxr-sr-x 2 root allusers 4096 Aug 25 21:44 a
> [root at bass1 ~]# touch a/b
> [root at bass1 ~]# ll a/b
> -rw-r--r-- 1
> root allusers 0 Aug 25 21:44 a/b
>
>
> This test is done under glusterfs file system
> [root at bass1 home]# ll -d a
> drwxr-sr-x 2 root allusers 6 Aug 25 21:45 a
> [root at bass1 home]# touch a/b
> [root at bass1 home]# ll a/b
> -rw-r--r-- 1 root root 0 Aug 25 21:46 a/b
>
>
> As we see here, the file b should belong to group allusers.
>
> Doesn't gluster support this feature? which could make group sharing easy.
> I would rather I did something wrong, please tell me how to fix this
> problem.
>
>
> Best,
> Manhong
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
>
>


-- 
Raghavendra G

A centipede was happy quite, until a toad in fun,
Said, "Prey, which leg comes after which?",
This raised his doubts to such a pitch,
He fell flat into the ditch,
Not knowing how to run.
-Anonymous

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20080825/ce333d2f/attachment.html>