Coldfusion Jrun apache connector recvTimeout and retry.

Question

Hi,
       I have set a recvTimeout for the JRUN apache connector using the
  JRunConfig RecvTimeout 300
        I want this setting so long running pages will time out and return after 300 s with a timeout error and return the error back to the client.
However, I see that the timeout is not working and long running cfm pages continue to cycle without returning. Upon enabling the verbose option I see the following error logged for the long running page after the timeout occurs, but the page does not return anything and continues to wait on the client side.
What does this error mean: jrRecv [11] timeout but server reachable so recv again
Is the jrun apache connector retrying or waiting some more for the page response ? I dont want it to retry / wait beyond the recvTimeout that I have set. What can I do in the config or elsewhere so that the recvTimeout returns as a timeout to apache and cleint without recv again attempt that it appears to be doing?
Thanks in advance for any help on this.
[Thu Dec 15 15:29:48 2011] [notice] jrApache[3068: 20480]     ap_escape_uri(TimeoutTest.cfm) = TimeoutTest.cfm
[Thu Dec 15 15:29:48 2011] [notice] jrApache[3068: 20480]  jrun_trans: sub-request so DECLINED
[Thu Dec 15 15:29:48 2011] [notice] jrApache[3068: 20480]     getRealPath(TimeoutTest.cfm) = TimeoutTest.cfm
[Thu Dec 15 15:29:48 2011] [notice] jrApache[3068: 20480]  sent/avail/result 79/79/79
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]  read/len/total 4/4/4
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]  PROXY_GET_HEADER <- [11]
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]  read/len/total 4/4/4
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]  read/len/total 21/21/21
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]     HTTP_X_REQUESTED_WITH: <null>
[Thu Dec 15 15:29:53 2011] [notice] jrApache[3068: 20480]  sent/avail/result 4/4/4
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  67.206.10.197:51050 jrConnect [9]
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  sent/avail/result 4/4/4
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  67.206.10.197:51050 is reachable
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  sent/avail/result 4/4/4
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  67.206.10.197:51050 is reachable
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  67.206.10.197:51050 returned socket [9] to pool
[Thu Dec 15 15:34:53 2011] [notice] jrApache[3068: 20480]  jrRecv [11] timeout but server reachable so recv again
[Thu Dec 15 15:36:33 2011] [notice] jrApache[3068: 20480]  read/len/total 4/4/4
[Thu Dec 15 15:36:33 2011] [notice] jrApache[3068: 20480]  PROXY_GET_REALPATH <- [11]
[Thu Dec 15 15:36:33 2011] [notice] jrApache[3068: 20480]  read/len/total 4/4/4

Answer 1

sunilr1 wrote:
Hi,
       I have set a recvTimeout for the JRUN apache connector using the
JRunConfig RecvTimeout 300
        I want this setting so long running pages will time out and return after 300 s with a timeout error and return the error back to the client.
That is not the setting to time out long-running pages and then return a message. Why would you want to control request timeouts from JRun anyway? I don't think it's a good idea.
RecvTimeout determines how long to wait before timing out a socket receive call on a JRun server. The receive call is generally a synchronous call on a socket on a JRun server. It will throw an exception if the time out period is exceeded. In other words,  a long-running page does not necessarily imply a long wait on a socket receive call.