As it stands right now, two messages are logged for each data-product
that's inserted into the product-queue whose creation-time is far enough
into the future (about 30 seconds) to cause problems. I could log fewer
messages, but at the risk of not identifying origination-sites with bad
I'll see if I can reduce the number without incurring the risk.
Michael Dross wrote:
I think that now with the ease in which we are able to sync clocks with
ntpd and other utilities, logging is
perfectly acceptable. If the log files rotate as they should, then
filling up of disk shouldn't be an issue.
I would only spew a incorrect time error in the log file at the startup
and every hour or so, not every minute.
Thanks for your work Steve,