I have not yet tried disabling KeepAlive. For our IPlanet servers we have always had KeepAlive turned ON. Also support recommended I make sure it is ON. I guess I can try turning it off just to see what happens, but I imagine there is a performance hit leaving it like that. I have tried increasing the KeepAlive timeout, and increasing the WLSocketTimeoutSecs and WLIOTimeoutSecs, but that did not help. I also have not fiddled with ConnectRetrySecs or ConnectTimeoutsecs yet. I didn't think would those were applicable here, since I increasing the sleep timeout would not be applicable. Also our application has not been returning any 503 Service Unavailable responses. -Joe -----Original Message----- From: Mike Rumph [mailto:mike.rumph@xxxxxxxxxx] Sent: Friday, July 01, 2016 10:03 AM To: users@xxxxxxxxxxxxxxxx Subject: Re: Issues migrating Weblogic proxies from Sun One 6.1 to Apache 2.4 Hello Joe, You can see the parameters that are available from the WLS plug-ins here: - https://docs.oracle.com/middleware/1213/webtier/PLGWL/plugin_params.htm#PLGWL475 The equivalent to what Yann was referring could be KeepAliveEnabled and KeepAliveSecs. But ConnectTimeoutSecs, ConnectRetrySecs, WLSocketTimeoutSecs and WLIOTimeoutSecs might also be useful. Thanks, Mike On 7/1/2016 3:06 AM, Joe Muller wrote: > As far as we can see from fiddler and the Apache logs the request is fully sent. I assume those parameters you mentioned are for mod_proxy ? We are not using mod_proxy for this, we are using the Weblogic plug-in for Apache 2.4. > > -----Original Message----- > From: Yann Ylavic [mailto:ylavic.dev@xxxxxxxxx] > Sent: Friday, July 01, 2016 3:42 AM > To: users@xxxxxxxxxxxxxxxx > Subject: Re: Issues migrating Weblogic proxies from Sun > One 6.1 to Apache 2.4 > > On Fri, Jul 1, 2016 at 12:49 AM, Joe Muller <jmuller@xxxxxxxxxxx> wrote: >> One interesting thing we did find on a packet capture on the back end between the proxy server and the application server is that 3 seconds after the proxy server apparently closed the port, the Weblogic application server tried to send data. It almost like Weblogic did not realize the port was closed. Very odd. > Is the HTTP request fully sent (proxy) and responded (backend) when this happens? (I try to understand why a backend/application would send data on its own). > > Otherwise, on a general rule, did you take care that the TTL of proxy connection (to the backend) is lower than the KeepAliveTimeout configured on the backend? > This can be done by using the ProxyPass/BalancerMember parameter ttl=... > But before that, does disabling keepalive on the proxy help already? > (ProxyPass' disablereuse=on or SetEnv proxy-nokeepalive) > > Regards, > Yann. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx > For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx > For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|