"Daniel Dang" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] > Bill, > > As your comment earlier I did open a bug here is a link: > http://issues.apache.org/bugzilla/show_bug.cgi?id=42608 > I don't know It's a right place to open this DR and those developers > have time to do anything with it. > > apache-tomcat is a new baby to me and I am not familiar with mod_jk > source. I am using apache-tomcat-5.5.16 and mod_proxy, mod_proxy_ajp > with server version Apache/2.2.3 >
After a brief look at mod_proxy_ajp, it looks like Tomcat should work with it (but mod_jk has issues). It would still be helpful if you could configure DEBUG logging for the category org.apache.jk (as in http://tomcat.apache.org/tomcat-5.5-doc/logging.html). This would let us see what mod_proxy_ajp is sending to Tomcat (which I very strongly suspect is wrong). You could also try asking on [EMAIL PROTECTED] to see if anyone there knows. > I have look at the latest version apache-tomcat-6 the > HandlerRequest.java in the decodeHeader method it looks very much the > same to apache-tomcat-5.5.16 version they still calling > MessageBytes.getInt(*) instead of getLong(). The getLong() method is > there and available in MessageBytes class. > I checked in a fix yesterday to the 5.5 branch to fix that. I'll port it to the 6.0 branch when I get a chance (there hasn't been an annouced release, so there is still time). You could get the latest from SVN and test that (you'll need to upgrade anyway to get Mark's fix). Also, 5.5.24 should be coming out soon, so you could just wait for that. However, if my guess is right, then the problem is with httpd sending a bad value for the Content-Length header, so you would have to report it over there and get it fixed. I don't have enough spare cycles to test this myself, but I'm sure that between you and Chris it will get nailed down. > If we modified anything with mod_jk*, I don't really know it will fix > the problem without modifying the tomcat? But whatever ways is work so > that I can able to send more then 3GB is the goal and I would like to > thanks to developers people like you guys. > > Cheers > > --Daniel Dang > > > > On Sun, 2007-06-10 at 13:16 -0700, Bill Barker wrote: >> "Christopher Schultz" <[EMAIL PROTECTED]> wrote in message >> news:[EMAIL PROTECTED] >> > -----BEGIN PGP SIGNED MESSAGE----- >> > Hash: SHA1 >> > >> > Mark, >> > >> > Mark Thomas wrote: >> >> Christopher Schultz wrote: >> >>> Since the Java part of the connector is bundled with Tomcat, is that >> >>> considered to be "part" of mod_jk (which I always thought of as the >> >>> c/c++ apache httpd module) or "part" of Tomcat? For instance, if I >> >>> were >> >>> to log a bug... where would it go? I don't see a proper product in >> >>> bugzilla (Connectors, Tomcat Connectors, mod_jk, Apache mod_jk, ... >> >>> none >> >>> of those exist). >> >> >> >> I'd use Connector:AJP but it doesn't really matter. >> > >> > That appears to be a component of the "Tomcat" product... I'm curious >> > why you make the distinction that this is /not/ a bug in Tomcat, but >> > rather in mod_jk. If the bug goes under the "Tomcat" product, then ... >> > ? >> > >> >> There isn't a lot of overlap at the moment between developers that work >> on >> mod_jk and those that work on the Java side. >> >> >>> I'd even be happy to provide a patch; I already made the mods in a >> >>> fresh >> >>> source copy and they compile... no testing, yet. >> >> >> >> Great. When you are happy with your patch, open a bugzilla entry and >> >> attach the patch. >> > >> > Okay, I'll get together with Bill and work on something. >> > >> >> mod_jk stores the content-length internally as an int. I think that >> older >> versions of mod_jk sent the internal version (sprintf or similar), and >> since >> the OP is using a very old version of Tomcat, he might be using an old >> mod_jk as well. With the latest versions of both, it looks like the >> header-parsing on the Java side will work (since the original header >> value >> is sent), but you won't actually get the request body because mod_jk >> thinks >> the content-length is negative. >> >> I haven't actually tested it, this is just from looking at the code. The >> mod_jk files to start with are apache-x.x/mod_jk.c and >> common/jk_ajp_common.c. HandlerRequest.java should probably be using >> longs >> as well, but it looks like it would recover eventually. >> > Thanks, >> > - -chris >> > -----BEGIN PGP SIGNATURE----- >> > Version: GnuPG v1.4.7 (MingW32) >> > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org >> > >> > iD8DBQFGbAl69CaO5/Lv0PARAuftAJ0euvzf+fCpMrA1x2PQ4dUQlFT6RQCgtQfO >> > dljS164b+JMDO+UYvdZjrpw= >> > =o8BI >> > -----END PGP SIGNATURE----- >> > >> > --------------------------------------------------------------------- >> > To start a new topic, e-mail: users@tomcat.apache.org >> > To unsubscribe, e-mail: [EMAIL PROTECTED] >> > For additional commands, e-mail: [EMAIL PROTECTED] >> > >> > >> >> >> >> >> --------------------------------------------------------------------- >> To start a new topic, e-mail: users@tomcat.apache.org >> To unsubscribe, e-mail: [EMAIL PROTECTED] >> For additional commands, e-mail: [EMAIL PROTECTED] > -- > > > --------------------------------------------------------------------- > To start a new topic, e-mail: users@tomcat.apache.org > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
