When my system is highly-loaded, the messeges are printed in mod_jk.log. [mod_jk.log] [Tue Dec 21 11:31:25.958 2010] [4744:3086915888] [warn] map_uri_to_worker_ext::jk_uri_worker_map.c (962): Uri * is invalid. Uri must start with / I assume the messages are caused by Internal Dummy Connection of Apache by the access_log. [access_log] ::1 - - [21/Dec/2010:11:31:25 +0900] "OPTIONS * HTTP/1.0" 200 - The OPTIONS request is valid, but mod_jk forejudges it as a WARNING. I made small patch when the uri of request matches '*', mod_jk doesn't print WARNING log. But I think it should be adjusted by the category of methods. (For example, if method is OPTIONS and uri is '*', debug message would output.) Best regards,
Created attachment 26438 [details] ignore the request which uri matches "*"
Please see http://markmail.org/thread/gmqxxlcoop6ijuoj for discussion on the user list. I'm not convinced this is a mod_jk problem, yet.
Since this is well-known behaviour of Apache, it would be nice if mod_jk would tolerate those dummy requests instead of filling the log file with warnings. In this special case there's nothing to warn about.
It seems like a request for "*" should never be mapped to mod_jk in the first place. Does "JkMount /*" match a bare "*" (with no leading /)?
Without support for OPTIONS * HTTP/1.1 - it's going tot be impossible to discern what additional methods are exposed by Tomcat, e.g. DAV directives. Consider this as a solution between Tomcat and mod_jk is sought.
Please use the latest release (1.2.35)
This message is still occurring in 1.2.40. What was the resolution on this? This message is filling the jk log too much.
The change is in http://svn.apache.org/viewvc?view=revision&revision=1598864 This has not been released yet but will be part of version 1.2.41. Until that version will be released (I plan to start the release cycle in a few weeks), you can include that simple patch if you do your own build. Regards, Rainer
*** Bug 57405 has been marked as a duplicate of this bug. ***
(In reply to Rainer Jung from comment #9) > *** Bug 57405 has been marked as a duplicate of this bug. *** Thanks for finding this... I was positive it was in BZ somewhere but couldn't find it because it was a) filed using different description and b) already fixed but c) not yet released.