autofs and namespaces

autofs and namespaces

[Tim Hockin]

For lack of clear insight into this, I thought I might put this out for 
opinions.

We're examining autofs and issues customers are having with it.  One of 
the issues that came up during discussions of things we can do is 
namespaces.

Does anyone have any ideas how namespaces and autofs ought to play 
together?  There are some obvious answers, but it seems to me that they 
might not be correct.

Anyone out there want to throw some ideas out?  We can just ignore it, 
but that isn't exactly nice.  What we really need is some guidance on 
what kind of sematics are "correct".

Tim


-- 
Tim Hockin
Systems Software Engineer
Sun Microsystems, Linux Kernel Engineering
thockin@sun.com

Re: [autofs] autofs and namespaces

[Ian Kent]

Hi Tim,

Fact is that I am just now starting to look at 2.6 and the issues relating 
to it wrt autofs.

Perhaps you (or someone else) could help by giving me a 5 cent tour of the 
issues as you see them.

On Wed, 20 Aug 2003, Tim Hockin wrote:

> For lack of clear insight into this, I thought I might put this out for 
> opinions.
> 
> We're examining autofs and issues customers are having with it.  One of 
> the issues that came up during discussions of things we can do is 
> namespaces.

Customers are having trouble with it.

One of them would be me.

As you have seen on the autofs list I have been working to improve Sun 
<-> Linux automount integration for some time now and while I don't have a 
complete solution I have made significant progress, in an operational 
sense.

With luck and effort I expect to have a release available soonish.

> 
> Does anyone have any ideas how namespaces and autofs ought to play 
> together?  There are some obvious answers, but it seems to me that they 
> might not be correct.

This sounds very interesting but I am deep into getting an operational 
release for the 2.4 kernel right now. Anything I can do, at least for a 
while, will be a cursory survey.

> 
> Anyone out there want to throw some ideas out?  We can just ignore it, 
> but that isn't exactly nice.  What we really need is some guidance on 
> what kind of sematics are "correct".

Don't want to ignore it but ....

-- 

   ,-._|\    Ian Kent
  /      \   Perth, Western Australia
  *_.--._/   E-mail: raven@themaw.net
        v    Web: http://themaw.net/