|
At 10.07 28/08/00 -0500, you wrote:
Hi Phil!
Umm submitting it as a feature/bug report on bugzilla?
http://bugzilla.redhat.com/bugzilla/enter_bug.cgi
Well, it didn't look like a bug to me... So I didn't want to offend anybody
posting on bugzilla! :)
That wasn't a 'feature' I thought sane people would want
however, I'll go off and make the necessary changes in the near future
Thank you VERY much! :) You're going to solve me a lot of troubles!
Features only get added when people suggest them, I get precious little
feedback
Well, this ML is for suggestion too, isn't it?
so your suggestion has become a requirement that will be added soon.
Though you
can
antagonise me and drop it in bugzilla (hint hint!) With luck that looks like
something I *might*
be able to code up buy middle next week.
Well, I'll wait for your precious "feature update", I've tried to write the
code for myself on top of piranha sources, but it was not as easy as I was
supposing...
Thank you very much for your WONDERFUL support, I'm looking forward for
good news! :)
Pietro Ravasio
P.S.: I suggest to treat each server/port couple as a single real server,
this way you can dinamically add/remove a bunch of real server maybe
running on the same IP (I'm going to run at least 20 application gateways
connecting to the main Database Server on the same phisical server! (I
don't have time to rewrite my application gateway to let it use threads,
and using LVS with multiple ports I can avoid queing a lot of clients'
connection requests adding parallelism (and using a few PCs)).
P.P.S.: another hint: why don't you add failover policies (start-stop
commands (script or whatever)) to be executed on a remote machine on a
per-real server basis on a "service failed" event? This way if an LVS
monitorized server falls down you can start a script that tries to start it
up again! (You certainly have already understood that my application
gateway is not inetd-driven! :PP )
|
