On Fri, 15 Jun 2001, ElephantChild wrote:
> of (.7, .6), the next (.6, .7) , then (.7, .6), etc. (IOW, *both
> addresses* are present in *all* answers, and the round robin just
> changes their *order* in the answer). So if both are up, clients will
> use both, and will hopefully spread their requests evenly. However if
> say, .6 goes down, the clients getting it first in the reply will have
> to try it first (perhaps getting an unreachable, perhaps timing out)
> before they try .7, but thet *do* have both addresses to try. (I
Indeed, after looking you are right. And yet I've never seen /anything/
actually /not/ simply time out using the first IP in the list it receives
if that server was down. At least not http clients. I cant recall having
tried it for anything else. (bind8 as the server) A reload/refresh
queries again, and gets the next server.
> - Or do the clients you have in mind ignore addresses after the first?
I suppose it would be that.
(there has often been discussion about this on isp-tech, and I dont seem
to be the only one that has seen it work this way.)
~shrug~
...david
---
david raistrick (deep in the south georgia woods)
[EMAIL PROTECTED]
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=8763&t=8308
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]