|
Sebastien,
Thanks for your reply, here some feedback.
--- Sébastien Bonnet <Sebastien.Bonnet@xxxxxxxxxxx>
> You're mixing several things. There's no "ipvsadm"
> monitoring. IPVS is
> *just* a scheduler, it doesn't care about service
> monitoring. This is
> acutally the role of nanny in piranha (can't
> remember the name of the
> equivalent of linuxvirtualserver.org).
Right, for those ports I have not explicitly told to
use my shell script or a string of chars to test the
port, then when I check "ps [-ef|-afx]" output, I see
the nanny daemons use ipvsadm with some parameters for
each of those ports. As you say, I suppose it is
using ipvsadm for scheduling those ports load
balancing.
> > TEST=`printf "quit\n" | telnet $1 25 2>/dev/null
> |grep
> > -i "connected"|wc -l` > /dev/null 2>&1
>
> Oh god ! Telnet isn't the right tool here because it
> takes ages before a
> timeout occurs.
Based on me actually testing this script, I can tell
you it DOES NOT take 10s of seconds to come back. It
comes back for all intents and purposes, instantly
with either an OK or a FAIL. Make port 25 to be $2
in the script and test non-existant ports too, and it
comes back instantly.
When you RS is working fine, your
> test works OK. When
> SMTP is down, your script won't return before
> several tens of seconds :(
Sorry guess I am dense - what is "RS"? Anyway I have
not seen a case where this script waited 10s of secs
to answer.
> There's something call nc (for netcat) which might
> be way more suitable.
Thanks, I will check it out.
Thanks
Peter
__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/
|
