Mailing-List: contact users-help@ace.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@ace.apache.org
MIME-Version: 1.0
In-Reply-To: 
 <CACtdb3JXdpDYdO2suRPN3W-XJ9iM21S8kvFpYJwv4MbXLUnbDQ@mail.gmail.com>
References: 
 <CABRbBKHtgaTivM9mfDcs6Lf+JP_P1A_z=CgkfyXTDTKePibK=Q@mail.gmail.com>
	<CACtdb3JXdpDYdO2suRPN3W-XJ9iM21S8kvFpYJwv4MbXLUnbDQ@mail.gmail.com>
Date: Thu, 1 Oct 2015 00:57:44 -0300
Message-ID: 
 <CABRbBKHxwSeppzMt9L+_rVcYUU1LM4fC5g1KygpVWVsVtddPDA@mail.gmail.com>
Subject: Re: Performance issues, logging, and relay server setup
From: "Robert M. Mather" <robert.mather.rmm@gmail.com>
To: users@ace.apache.org
Content-Type: multipart/alternative; boundary=001a11409b68a774920521030c46

--001a11409b68a774920521030c46
Content-Type: text/plain; charset=UTF-8

On Sep 2, 2015 8:57 PM, "Karl Pietrzak" <kap4020@gmail.com> wrote:
>
> On Wed, Sep 2, 2015 at 5:12 PM, Robert M. Mather <
> robert.mather.rmm@gmail.com> wrote:
>
> > We have about 200 targets running off ACE in production now, and as we
add
> > more, we're noticing a steady decline in performance of the web UI,
> > especially the initial login. Also, occasionally the server seems to get
> > caught in a loop or something and starts eating up CPU until we kill and
> > restart it. After restart, it always seems to perform a bit better for a
> > while, but the cycle continues.
> >
> > Is there some way we can get debug/logging output from the ACE server?
That
> > way we'd have a better sense of what's going on and we could report back
> > with better information.
> >
> > Also, due to the performance issues and our need to continue expanding,
> > we're considering setting up a relay server. We're running on the
> > 'server-allinone' distribution at the moment, would we have to switch to
> > the "componentized" version in order to configure for relay?
> >
> > Any other recommendations?
> >
> > Not sure if it matters, but our targets sync with ACE every 7 minutes.
Is
> > it possible the cost of that ping is starting to add up?
> >
> > Thanks!
> >
>
> Is there a way for you to run on profiling (are you familiar with the
> concept)?  Even a little profiling data would be a huge help.

Do you mean jvm profiling? I haven't worked with that, but could give in a
try. By the way, our ace server crashed the other day for no apparent
reason. It would be very helpful to capture log4j (or whatever logger ace
uses) output to a file.

Sorry for the delay in my reply, I didn't see your message when it arrived
in by inbox for some reason.
>
>
> --
> Karl

--001a11409b68a774920521030c46--