We recently had problems building resin-pro-3.1.8 for Apache and
applied the solution in 0003698 of removing line CFLAGS="$CFLAGS from
the configure file. That change allowed a successful build in a test
Using this same method on our new production server results in
"Service Temporarily Unavailable" through Apache, but success direct
to Resin. I thought this was related to the previous build bug, but I
have gone back to 3.1.3 and have the same issue.
The new server is CentOS rel 5.2 with httpd/devel 2.2.3-11, libgcc/gcc
4.1.2 and built with: ./configure --prefix=/opt/resin-318pro
--enable-jni –with-apxs=/usr/sbin/apxs Caucho-status returns a
positive "Configuration Cluster" but "Unconfigured Default Virtual
Changing configure arguments, e.g. --with-apache-cflags="-DLINUX=2
-D_REENTRANT -D_GNU_SOURCE -pthread" --enable-debug –enable-static=no
has the same results, as does changing Apache conf between
ResinConfigServer and CauchoHost. Apache and Resin confs have been
edited from basic post install to fully mimic our production confs.
All variations have given the same result.
Has anyone else seen this behavior with new builds of CentOS? Can
someone suggest a method to isolate the problem to the mod_caucho
build or conf file.
resin-interest mailing list