于 2008年10月22日 22:58, Jason King 写道:
> On Wed, Oct 22, 2008 at 7:08 AM, yifan xu <[EMAIL PROTECTED] 
> <mailto:[EMAIL PROTECTED]>> wrote:
> 
>     The VRRP project team invites you to review our current design. The
>     document could
>     be found at:
> 
>     http://www.opensolaris.org/os/project/vrrp/vrrp_design.pdf
> 
>     It would be appreciated if you could provide any comments on the
>     networking-discuss
>     forum no later than the 7th, November.
> 
> 
>     Cheers,
>     Yifan
>     --
>     This message posted from opensolaris.org <http://opensolaris.org>
>     _______________________________________________
>     networking-discuss mailing list
>     [email protected]
>     <mailto:[email protected]>
> 
> 

Jason,

Thanks for your quick respnse.

> A couple of (hopefully not too stupid) questions:

Definitely not. :)

> Is there a reason the configuration cannot be managed entirely via smf?  
> Looking at it, it seems like it would fit well.

The configuration is too complex for smf to manage; especially, some 
properties are nested within other properties. Please see page 16/17 for 
an example configuration file.

> It seems doors are the preferred IPC mechanism for such things on 
> *solaris, any reason unix sockets are being used instead?  

Doors depend heavily on threads, which we'd like to avoid. And we're 
more familiar with unix domain sockets. :)

> Also, 
> shouldn't it (unix socket or door file) go in /var/run instead of /tmp?

This sounds reasonable. Will change.

--
Huafeng
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to