On Mon, Aug 30, 2004 at 02:36:47AM -0700, Ram wrote:
> The client has no time waiting. he wants to send "some important"
> messages to "n" number of webservers. because the client has to send
> messages to "n" number of webservers. He cannot wait for one of them
> because others will be delayed.

Aah, that's what you're getting at!

IMHO you are attempting the wrong approach here. Look at the curl-multi
interface; using it, you will be able to send the same request to multiple
webservers _in_parallel_, which will be _way_ faster than sending many
requests sequentially.

You really should let libcurl receive the server responses. If you think
about it, it might even be useful to analyse them in the tool you're
writing so it can return a meaningful status message like "x out of y
servers returned an OK status to the reboot command, z servers returned an
error, n servers were not reachable".

Cheers,

  Richard

-- 
  __   _
  |_) /|  Richard Atterer     |  GnuPG key:
  | \/¯|  http://atterer.net  |  0x888354F7
  ¯ '` ¯