[MTOS-dev] Roadmap?

[Timothy Appnel]

Yeah this is really denting MT and my creditability with a couple of
clients currently. Telling them the new system I am rolling out (or
have) has a memory leak that will eventually crash their server was
bad enough. Telling them the fix is to restart the server/daemon
periodically I hope you restart it frequently enough --- well let's
just say it doesn't inspire confidence. Spinning that like its fine is
bad for business.

It's a bit bothersome to see that effort is going into refactoring the
search, putting in a performance monitoring framework and modifying
the template listing screen (what happened to performance only
release?) and yet nothing has been communicated. I'm not even sure
Hirotaka's patches have been checked in.

<tim/>

On 2/28/08, Jay Allen <jay at endevver.com> wrote:
> Bump.  I'm actually curious about the answer below.  Would love to
>  hear thoughts on the issue from anyone who has them.
>
>  Jay
>
>  On Sat, Feb 23, 2008 at 2:07 PM, Jay Allen <jay at endevver.com> wrote:
>  > Can one of you clear something up for me?  While memory leaks are bad
>  >  and certainly need to be addressed, are they something that can't be
>  >  mitigated by modifying your process lifetimes?  Are they significantly
>  >  worse than those experienced under MTE 1.53 with FastCGI and
>  >  RebuildQueue in daemon mode?
>  >
>  >  Again, I'm not saying that they shouldn't be fixed forthwith; Only
>  >  that I've been successful in recycling RQ and FastCGI periodically and
>  >  am wondering if even that is no longer possible.
>  >
>  >  Jay
>


-- 
Timothy Appnel
Appnel Solutions
http://appnel.com/