[ic] Verisign, double, tripe charges, orders not going through IC

[interchange-users@interchange.redhat.com]

Quoting Ron Phipps (rphipps@reliant-solutions.com):
> > I am a bit disappointed in the (well-intentioned, I am sure)
> > finger-pointing toward IC, though, given your apparent old Perl and
> > Linux. You would be well-advised to update to Perl 5.6.0 -- the latest
> > stable Perl is always assumed if you don't mention it. I am surprised
> > that it is the first I noticed that, and am surprised that you would
> > deploy a system with a two-year-old-plus Perl. If you are using a
> > two-year-old Linux kernel as well I would advise updating that as
> well.
> 
> I apologize for the finger pointing Mike.  I'm aware that other things
> can cause problems on the system since the site depends on so many
> different modules to perform its functions.  It is definitely my fault
> for not realizing to update the Perl to the latest and greatest as a
> precaution to installing a new system which relies on Perl.  We have had
> this site deployed for a year now and it has been rock solid from 4.0 ->
> 4.6.1 despite the old perl. 
> 
> > Remember all bugs are not Interchange. 8-\ If you didn't know,
> > thousands of bugs have been fixed in Perl and Linux in the last two
> > years.
> 
> I know this :)  And once again I apologize if you took my message as an
> attack on IC.  You know that I'm very thankful for the software that you
> have provided and do not take it for granted.  I got tied up in the
> moment of finding a quick fix instead of looking at the situation from
> further away.  Those things happen when a launch goes bad and sales are
> being lost ;)
> 

I didn't see it as an attack -- but it is easy to lose site of how much
a system depends on all its parts. I was trying to bring that out, and
trying to emphasize how much Interchange depends on its underlying
platform. Did you mention a multi-CPU system? I am not sure. If you have
a multi-CPU system, then that applies 10 times over.

All that being said, I am certainly not asserting that there could
not possibly be a problem with Interchange 4.8. I know better than
that. I just think it is more likely to be the kernel or Perl.

After racking my brain, the only thing I can think of, is that
Interchange 4.8.x does that 4.6 didn't that might conceivably affect
this is to use the reset() function in Perl. There may have been some
bugs with that function fixed in Perl 5.6.

> 
> > The fact that the problem only appears at busy times means it can
> > be anything, and that it is less likely to be Interchange than it
> > is overall system.
> 
> Point noted, we will be testing and deploying the new perl and will then
> readdress IC 4.8.x and see if the problem is tied to the old perl.  I
> have watched the load averages and memory usage when the problem occurs
> and it is independent of how busy the site is.
> 

[snip] 

> I hope you see my true intentions (to resolve the situation quickly) and
> look over my frustration which was expressed in my message.
> 

Good luck. I understand the frustration, believe me. I still remember
flaky problems from my days of bringing high-volume sites online.

-- 
Red Hat, Inc., 3005 Nichols Rd., Hamilton, OH  45013
phone +1.513.523.7621      <mheins@redhat.com>

I am a great believer in luck, and I find that the harder I work
the more luck I have. -- Thomas Jefferson