2014-06-16 19:44 GMT+04:00 Martin Knoblauch <knobis...@gmail.com>:
> Hi,
>
> my environment is:
>
> OS SuSE SLES11/SP2 64-bit
> HTTP frontend: Apache2-2.2.21 (64-bit)
> mod_jk: 1.2.37
>
>
> Problem short: excessive connection timeouts with weird socket ids
> Problem long: we frontend 4 Tomcats (7.0.42) using Apache2(2.2.21) and
> mod_jk(1.2.37). We are seeing a heavy amount of connection timeouts when
> the tomcats are loaded. Surprisingly the logs show socket id "-1", which
> should not happen in my opinion, because most actions are guarded by the
> IS_VALID_SOCKET macro. Here is one example. The code emmitting the line
> should not be reached due to the check in line 1543:
>
> [Mon Jun 16 17:12:05.405 2014] [7233:140275337774848] [info]
> ajp_send_request::jk_ajp_common.c (1562): (lpsdm83-1) failed sending
> request, socket -1 prepost cping/cpong failure (errno=110)
>
>
> 1543: while (IS_VALID_SOCKET(ae->sd)) {
> int err = JK_FALSE;
> if (jk_is_socket_connected(ae->sd, l) == JK_FALSE) {
> ae->last_errno = errno;
> jk_log(l, JK_LOG_DEBUG,
> "(%s) failed sending request, "
> "socket %d is not connected any more (errno=%d)",
> ae->worker->name, ae->sd, ae->last_errno);
> ajp_abort_endpoint(ae, JK_FALSE, l);
> err = JK_TRUE;
> err_conn++;
> }
> if (ae->worker->prepost_timeout > 0 && !err) {
> /* handle cping/cpong if prepost_timeout is set
> * If the socket is disconnected no need to handle
> * the cping/cpong
> */
> if (ajp_handle_cping_cpong(ae,
> ae->worker->prepost_timeout, l) == JK_FALSE) {
> 1562: jk_log(l, JK_LOG_INFO,
> "(%s) failed sending request, "
> "socket %d prepost cping/cpong failure (errno=%d)",
> ae->worker->name, ae->sd, ae->last_errno);
> /* XXX: Is there any reason to try other
> * connections to the node if one of them fails
> * the cping/cpong heartbeat?
> * Tomcat can be either too busy or simply dead, so
> * there is a chance that all other connections would
> * fail as well.
> */
> err = JK_TRUE;
> err_cping++;
> }
> }
>
> So, before trying to tell me that I am missing workers.properties and
> complete logs, I want to check first whether the "-1" socket ids are to be
> expected here or whether there is some kind of bug involved.
>
The ajp_handle_cping_cpong(...) method that is called a few lines
earlier (line 1560) processes some failures by calling
ajp_abort_endpoint(...). That method closes the socket and sets it to
"-1".
Thus the socket is not "-1" by the start of the loop, but can become
"-1" in the process.
Best regards,
Konstantin Kolinko
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
