-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Moran,
Moran Ben-David wrote: >> Nobody uses the jni worker anymore. The jni worker was intended to be an >> embedded Tomcat instance running from the mod_jk module, and is only >> supported in very old versions of Tomcat (3.3 IIRC). > > Is there any particular reason why no one is using it? It seems that > with large loads of http requests and responses the in-process > interface is ideal. Most people just use Tomcat standalone if they are worried about the performance hit of putting httpd in front of Tomcat. Realistically, the performance drain on an application is not the time it takes to transfer data from the NIC to the app server: it's the time to decode SSL requests. > My assumption is that the jni worker not being used anymore has > something to do with apache not having good if any multi-threaded > handling for modules at the time. However, with the advent of > apache2's MPM worker tomcat's design (multiple threads for multiple > requests) should fit like a glove. > > Does this make sense? Or am I am I drawing up a neat fantasy rather > than a plausible story? It does make sense. I think it's just that nobody cares anymore. That feature was such a pain to support because nobody understood it well enough to use it properly. The potential for idiotic use was /way/ to high. People already have a super-hard time compiling their own mod_jk modules even though the process is as simple as "./configue --with-apxs=... && make && cp ...". Compiling a JNI module for a particular platform would cause even more confusion. There's even confusion when binaries are provided. It's just not worth it. Don't forget that this is a Linux-only (potential) optimization, and only on certain versions of the kernel. Given that so many OSs are supported by Apache httpd and mod_jk, it makes sense to use the standard APIs whenever possible. >> That's because the JNI options are not applicable to new versions. > > Is there a way to compare how tomcat would run in-process with apache2 > MPM? Do you know of any benchmarks in this direction? I know of no published benchmarks. You should easily be able to set up Tomcat 3.3 + mod_jk JNI versus Tomcat 5.5 and "regular" mod_jk and see what happens. My guess is that there won't be too much difference. - -chris -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHE2vB9CaO5/Lv0PARAiwKAKC+BG0SydV3EI5PX/aA0ry7GxrTlACbBVuR 8Zut98Xt/voW6JU7p81ApT4= =jn0p -----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]