httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mladen Turk <mt...@apache.org>
Subject Re: svn commit: r682369 - /httpd/httpd/trunk/modules/proxy/proxy_util.c
Date Mon, 04 Aug 2008 14:29:11 GMT
Ruediger Pluem wrote:
> 
> 
> On 08/04/2008 03:50 PM, Mladen Turk wrote:
>> rpluem@apache.org wrote:
>>> --- httpd/httpd/trunk/modules/proxy/proxy_util.c (original)
>>> +++ httpd/httpd/trunk/modules/proxy/proxy_util.c Mon Aug  4 05:53:50 
>>> 2008
>>> @@ -1380,7 +1380,6 @@
>>>      proxy_worker *worker = (proxy_worker *)theworker;
>>>      if (worker->cp->res) {
>>>          worker->cp->pool = NULL;
>>> -        apr_reslist_destroy(worker->cp->res);
>>>      }
>>>      return APR_SUCCESS;
>>>  }
>>>
>>
>> Well, you've beat me :)
>> The complete conn_pool_cleanup function can be omitted, and
>> there is no need for checking  if (conn->worker->cp->pool)
>> in the connection_destructor.
> 
> Might be. At the current state it makes it possible to keep
> trunk and 2.2.x very similar until all in apr land is sorted
> out. We can cleanup the remaining parts then.
> 

Well I was hopping to add an extra apr_reslist_maintain function
that would allow to close the expired resources out of the
require/release bounds. Right now in proxy the connection will
be release only if the particular mpm child process gets
request for a particular worker. It can take minutes or hours
on slow hit ratio for that to happen. So the ttl is more a 'guess'
rather then a firm fact.


Regards
-- 
^(TM)

Mime
View raw message