On 08.05.13 00:05, Tim Bunce wrote:
[...]
At this stage in the life of the DBI I think it's reasonable to assume
that there isn't a leak in the DBI itself. If there was then a lot of
people would be affected and complaining about it.

I think this assumption is heavily wrong. From my jobs I learned one
thing the last years: do it quick. To illustrate what I mean saying
that, see following exampple.

In a particular project we had the issue, that every monday morning
our web-interface was inoperable because of an invalid $dbh (which
was cached). The question from our manager was: can you say in 10
minutes why and how to fix and can you guarantee fix it until lunch?
Otherwise - implement a cron controlled web-server restart each
morning at 5.

Because of personal interests I debugged later into it (and fixed
it a few weeks later - "guaranteed") ...

What I'm trying to say - when administrators see there web-server
machines running out of memory, the implement a scheduled restart
in the pool of machines.

I really don't know business relying on web-services running
qualified monitoring with performance measurement (some do, but
don't know what they measure and finally one company really does).

Sorry for complaining ;)

Cheers
--
Jens Rehsack

Reply via email to