On Mon, 2 May 2005, Jeffrey Layton wrote:
On Mon, 2005-05-02 at 22:01 +0800, [EMAIL PROTECTED] wrote:On Mon, 2 May 2005, Jeffrey Layton wrote:
Here's the latest version of the replicated mounts patch that I've been working on. I sent this a couple of weeks ago and got no response.
I've been trying to get my direct mount stuff working.
Sounds good, just wanted to make sure it didn't slip thru the cracks. I also wanted to make sure I don't allow any regressions as I update my patch for new CVS versions.
Also, I've been doing some looking at what it will take to do the "subnet" locality check. The problem seems to be quite a bit more complicated than it seems at first glance. I'm trying to wrap my brain around netlink sockets as that seems to be the best way to get routing/netmask info out of the kernel.
Yep. Sounds good.
An interesting sideline to this might be to consider how this would apply to discovering hosts for a "-hosts" map type. Much later, but worth some thought.
We might also want to consider patching "mount" for that effort since the current mount command just picks the first address when given a hostname that resolves to more than one address.
Ouch.
If/when my other patch goes in, I'll start focusing on that.
It's a big change.
It would be good if we could get it right as the follow on overhead of subtle bugs is huge.
Ian
_______________________________________________ autofs mailing list [email protected] http://linux.kernel.org/mailman/listinfo/autofs
