On 28.07.2010 07:44, Mladen Turk wrote:
> On 07/27/2010 08:00 PM, Mark Thomas wrote:
>>>
>>> What happens if you disable the terminateAPR call?
>>
>> Thanks. That was it. I changed the point where that was called and that
>> fixed it although it did then highlight a huge delay (~30s) to shutdown
>> the connector. I've got to the bottom of that too and am currently
>> testing a patch.
>>
>
> One other reason for a potential core on shutdown is
> the accept call. I did however some nasty tricks to prevent that
> directly in native, but the bottom line is that acceptor
> should be broken before its pool gets destroyed.
>
> With APR 1.4.x there is a new pool pre-cleanup hook
> that would allow to effectively handle those situations,
> but it would also mean that native won't work on pre 1.4.x APR's
If it is important, we could use that for TC 7 using a newer tcnative
branch?
Regards,
Rainer
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
|