Coda File System

Re: modular clog + kerberos

From: <>
Date: Mon, 25 Jan 2010 13:26:35 +0100
Hi Don,

On Mon, Jan 25, 2010 at 03:28:54AM -0800, root wrote:
> Greetings all: 
> Is the following normal (not seeing /coda/coda.realm when doing ls /coda): 
> [root_at_sandbox3 ~]# ll /coda/coda.realm

By the way, what is "ll"? I may guess but guessing is dangerous. :)

> [root_at_sandbox3 ~]# ll -d /coda/coda.realm
> drwxr-xr-x 1 root 65534 2048 2010-01-15 06:55 /coda/coda.realm
> [root_at_sandbox3 ~]# ll /coda/
> total 0

> It looks ok, although with the "stock" coda I could look at /coda and see 
> /coda/coda.realm.  Is this "normal" or is something still broken? 

I would expect seeing the realm name in /coda but it shouldn't influence
anything if it does not stick there.

> I am working on implementing multiple coda volumes -- we need complete 
> isolation of data between environments/groupings of servers, and having a 
> volume for each seems the most thorough means of doing it.  Is there any 
> logistical problems or limitations to doing this?  Otherwise any insights 

(The word "servers" forced me to read the paragraph twice to understand
what you mean. As an example, "servers for different services like web"
would be much easier to interpret)

Be aware of automated and especially concurrent updates on Coda,
anything that makes updates must have the knowledge to cope with possible

Dividing the data into volumes is good.
See the wiki for limitations.

> on how to limit use of users to specific volumes would be greatly 
> appreciated. 

Note, you set ACLs on directories, not on volumes.

A volume is a quota unit, backup unit, concurrency serialization unit -
but not an "access rights" unit.

> And, of course, thanks again for the very active and knowledgable support
> community.  It is far and above what I had dared to hope for.

You are welcome.

Received on 2010-01-25 07:27:34