paul.nowak wrote: Matt, thanks for the comments. I made an error on the version of Plone. It's 2.5 Plone running on Zope 2.9x.
In regards to the additional products, we have a skin installed and we have a product that we had custom developed for us that connects to a PostgreSQL database. We've looked at slow PostgreSQL queries causing problems and have not been able to find an issue. We've also tested for the case where the PostgreSQL server is down and have not been able to create an issue. We therefor...
BURBANK, CA -- (Marketwire) -- 07/06/09 -- That's a bit of a joke, really; was there ever
a time when response time didn't matter? Back in the old days, roomfulls of
secretaries and file clerks were kept running to make sure executives were
provided with data from files in a timely manner. The whole idea of
computerizing data was to cut response time even more by removing the human
lag factor from the response equation; at the mere press of a button, the
data would be instantaneously available. No more screaming across rooms or
down halls if documents weren't procured quickly enough, and no more
employee reprimands or, at worst, pink slips issued for multiple
infractions.
But like an ironic twist out of a science fiction story, the computer era
brought its own form of non-response. Imagine the look of surprise on that
old-school executive's face when he finally got to press that proverbial
button -- and nothing happened? It is certain that even more screaming
ensued. Unfortunately, it's continued to occur frequently since then.
The culprit is file fragmentation. For applications or even the operating
system the story is the same: when files are fragmented, response time is
greatly slowed down. When fragmentation gets seriously bad, it goes from
slow response time to no response time. It's as bad as yelling down to a
file clerk that has fallen asleep or taken an unscheduled vacation.
Early on, system administrators did get wise to fragmentation and solutions
were evolved for it. First they were all manual, then they could be
scheduled. For many years, scheduled defragmentation did the job and
everyone seemed happy.
But now a new era has dawned, the era of the World Wide Web and global
business. The maintenance time windows that were previously available for
scheduling or manually running defrag have all but disappeared, while
systems remain up and running 24X7. In between the few defrag runs that can
be scheduled, fragmentation continues to build and cripple response time.
Today's executives don't remember back to the days of human file clerks,
but they certainly know that computer response time should be as close to
instant as possible. Right they are, and some sites are experiencing
fragmentation's effects simply because of choosing an outdated defrag
technology.
The only way to adequately deal with today's fragmentation issues is with a
fully automatic defrag solution, one that operates invisibly, in the
background. Fragmentation is always addressed and is never a problem. Only
otherwise-idle resources are used, so there is never a negative performance
effect on users, and best of all scheduling is never required.
Yes, response time still matters and always will. Make sure to choose a
defrag technology that will keep it maximized.