Stefano Bagnara wrote:
> What does the "Confirmed" word in the subject means?
As noted in the original e-mail, it means that I had it happen over the
course of a week in my production environment, but that I had not dug into
the heap to see what went wrong, nor have I had (nor did I expect to have)
any time at all to dig into the heap this week.
> What mysql drivers do you use? What mysql server do you use?
The same driver that we used to ship. And the memory leak isn't in the
database server process, which generally runs non-stop for anywhere from
months to over a year without problem. And non-JAMES JVMs in my production
environment that use the same driver have been running for months without
any noticable memory leak.
> I suspect the problem may be in a buggy mysql driver
Possibly, although from what little research I had time to do, it seems that
more bugs were introduced AFTER the version I am using than before. So it
could be a buggy driver, or perhaps some JAMES component matcher/mailet not
tested in our stock configuration.
> or simply you give not enough memory to james for your traffic.
That would be a no. The process grew slowly for a week until it blew up.
And I keep a very tight leash on the maximum message size permitted.
> Please never again post a "blocker" message like this without version
> informations and configuration if you plan to not be able to answer
> questions fast ;-)
Too bad. It wasn't as if I thought it was a good idea to indicate --- as I
packed to leave for the airport --- that I had run out of memory in a
production environment. We can ship if that's what everyone wants. I'll
dig into it when I have time, but this was not a good week time-wise for me.
I might have time to restart the JVM and run with hprof this weekend when I
can babysit the server.
--- Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
