modcluster / mod_cluster

mod_cluster is an intelligent native Apache httpd-based and pure-Java Undertow-based load-balancer

mod_cluster

Project mod_cluster is an intelligent load balancer. It uses a communication channel to forward requests from a reverse proxy server to one of a set of application server nodes. Unlike mod_jk and mod_proxy, mod_cluster leverages an additional connection between the application server nodes and the reverse proxy to transmit server-side load balance factors and lifecycle events back to the proxy. This additional feedback channel allows mod_cluster to offer a level of intelligence and granularity not found in other load balancing solutions. There are currently two reverse proxy implementations: a native Apache HTTP Server implementation and a pure Java Undertow-based implementation.

Project mod_cluster boasts the following advantages over other httpd-based load balancers:

• Dynamic configuration of httpd workers
• Server-side load balance factor calculation
• Fine grained web-app lifecycle control
• AJP is optional

https://www.modcluster.io

Installation Instructions

JBoss AS 7/WildFly 8 (or newer)

These versions already ship with bundled mod_cluster. It is configured via mod_cluster subsystem.

JBoss AS 6

This is the first version of AS that bundled mod_cluster, the configuration is located in /server/<profile>/deploy/mod_cluster.sar/META-INF directory.

Tomcat 8.5 (or newer)

Distribution archives are provided for each Tomcat version.

1. Obtain the distribution archive corresponding to the intended Tomcat version by either downloading from the project website or if building from source located in dist/target/ directory.

2. Download and unzip or untar the distribution archive and navigate to the extracted directory.

3. Copy the lib/ directory to the Tomcat installation directory adding jars to its lib/ directory. If upgrading from a different version, it is necessary to remove all jars copied previously.

4. Modify server.xml within the conf directory and add the mod_cluster listener as documented here. The minimal listener configuration is as follows:

   <Listener className="org.jboss.modcluster.container.tomcat.ModClusterListener" connectorPort="8080" advertiseInterfaceName="lo0"/>

Project Structure

Project is split up into multiple modules:

core (contains the implementation of container-independent core mod_cluster concepts)
container
  spi (SPI classes for container integrations, has no dependencies on a specific web container)
  tomcat-8.5 (Tomcat 8.5 container implementation)
  tomcat-9.0 (Tomcat 9.0 container implementation)
  tomcat-10.1 (Tomcat 10.1 container implementation)
load-spi (SPI classes for load metric computation)

Source Code

Source code for the mod_cluster project is located on GitHub:

https://github.com/modcluster/mod_cluster

Building

Servlet Container Modules

When building from source, first ensure that Maven version 3.2.5 or newer (run mvn -version) and JDK 11 or newer (run java -version) are installed. The following command builds modules for all containers:

mvn clean install

Distribution files for Tomcat will be built in the dist/target/ directory.

Code Coverage Report

This project currently supports JaCoCo to generate a code coverage report bound to verify goal:

mvn clean verify -P coverage

The resulting report can be viewed by opening code-coverage/target/site/jacoco-aggregate/index.html.

Reporting Issues

Project mod_cluster uses Red Hat Jira issue tracker under MODCLUSTER project:

https://issues.redhat.com/browse/MODCLUSTER

License

• Apache License 2.0

mod_cluster's Issues

wiki:pull-request Apache HTTP Server configuration initial (2014-11-19 13:46:04 +0100)

Hi guys, my

git request-pull 261a55d0 https://github.com/Karm/mod_cluster.wiki.git 261a55d0

generated this lovely message for you:

The following changes since commit 261a55d065b496baffb596ae2727dc43533b5006:

  Apache HTTP Server configuration initial (2014-11-19 13:46:04 +0100)

are available in the git repository at:

  https://github.com/Karm/mod_cluster.wiki.git master

for you to fetch changes up to 261a55d065b496baffb596ae2727dc43533b5006:

  Apache HTTP Server configuration initial (2014-11-19 13:46:04 +0100)

----------------------------------------------------------------

TLS certificate is being provisioned. This may take up to 15 minutes to complete.

Ive had it saying TLS certificate is being provisioned for the past hour or so. I deleted and reset it and no use. Is this an error, can the mods do anything from their end or anything i should do from my end.

error: /opt/jboss/httpd/lib/libapr-1.so.0: undefined symbol: dlopen

I do this:
tar -xzvf mod_cluster-1.2.6.Final-linux2-x64-ssl.tar.gz
/opt/jboss/httpd/sbin/apachectl start
the error is
/opt/jboss/httpd/sbin/httpd: symbol lookup error: /opt/jboss/httpd/lib/libapr-1.so.0: undefined symbol: dlopen

why? please help me! many thanks!

jvm-route encryption

Hello guys.

First of all, this is a feature request and not a bug.

I would like to "obfuscate" the jvmRoute so that an external attacker cannot "guess" the topology of my internal infrastructure.
The "strong" way would be to have a symmetrical cipher with a configurable key.
mod_cluster could then cipher the jsessionid before exposing it to the external world, and decipher it to recover the jvmRoute and properly redirect the request.
But I guess that this would have very undesirable consequences on performance.
The "weak" way would be just obfuscate, i.e. let's say that the jsessionid is alea + '.' + jvmRoute. We could take a part of the alea to alter the jvmroute in a reversible way (XORing for instance).
Anyhow, the expected effect would be that the jvmroute would be externally different for each and every request.

Unfortunately, I have close to no C skills, hence I cannot make this myself.

(as a side note, coming from mod_jk, I'm quite impressed by the features mod_cluster offers! Thanks for the good work :) )

Prometheus metrics for mod_cluster_manager

Has there been any thought about having mod_cluster expose prometheus metrics for scraping context status over time instead of relying on soley on mod_cluster_manager? Gets a bit much to parse through when we have many apps over many nodes, currently for many projects I see over 100 contexts total

Cannot Compile HTTPD mod_proxy_cluster module

When running make inside the latest release of 1.3.11.Final the native/mod_proxy_cluster
Fails with the below error on Linux
After running
./buildconf
./configure --with-apxs=/usr/sbin/apxs
./make
/usr/lib64/apr-1/build/libtool --silent --mode=compile gcc -pthread -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic -Wformat-security -fno-strict-aliasing -DLINUX=2 -D_REENTRANT -D_GNU_SOURCE -I/usr/include/httpd -I. -I/usr/include/apr-1 -I../include -prefer-pic -c mod_proxy_cluster.c && touch mod_proxy_cluster.slo
mod_proxy_cluster.c: In function ‘init_conn_pool’:
mod_proxy_cluster.c:237: warning: unused variable ‘cp’
mod_proxy_cluster.c: In function ‘internal_find_best_byrequests’:
mod_proxy_cluster.c:2393: error: ‘pptr’ undeclared (first use in this function)
mod_proxy_cluster.c:2393: error: (Each undeclared identifier is reported only once
mod_proxy_cluster.c:2393: error: for each function it appears in.)
mod_proxy_cluster.c:2395: error: ‘proxy_worker_shared’ undeclared (first use in this function)
mod_proxy_cluster.c:2395: error: expected expression before ‘)’ token
mod_proxy_cluster.c: In function ‘proxy_cluster_post_config’:
mod_proxy_cluster.c:2899: error: ‘proxy_worker_shared’ undeclared (first use in this function)
make: *** [mod_proxy_cluster.slo] Error 1

The above process works fine with 1.3.10.Final

Versions
$ apachectl -v
Server version: Apache/2.2.15 (Unix)
Server built: Jun 19 2018 15:45:13

httpd-2.2.15-69.el6.centos.x86_64
httpd-devel-2.2.15-69.el6.centos.x86_64
gcc-4.4.7-23.el6.x86_64
make-3.81-23.el6.x86_64

"mvn -P dist package" gives "Unresolveable build extension"

pi@raspberrypi ~/mod_cluster/mod_cluster $ ~/apache-maven-3.2.1/bin/mvn -P dist package
[INFO] Scanning for projects...
Downloading: http://repository.jboss.org/maven2/org/jboss/jbossorg-docbook-xslt/1.1.0/jbossorg-docbook-xslt-1.1.0.pom
Downloading: http://repository.jboss.org/maven2/org/jboss/jbossorg-jdocbook-style/1.1.0/jbossorg-jdocbook-style-1.1.0.pom
Downloading: http://repository.jboss.org/maven2/net/sf/docbook/docbook/1.72.0/docbook-1.72.0.pom
Downloading: http://repository.jboss.org/maven2/javax/media/jai-core/1.1.3/jai-core-1.1.3.pom
Downloading: http://repository.jboss.org/maven2/com/sun/media/jai-codec/1.1.3/jai-codec-1.1.3.pom
[ERROR] The build could not read 1 project -> [Help 1]
[ERROR]
[ERROR] The project org.jboss:User-Guide-en:1.3.1.Final-SNAPSHOT (/home/pi/mod_cluster/mod_cluster/docs/userguide/pom.xml) has 2 errors
[ERROR] Unresolveable build extension: Plugin org.jboss.maven.plugins:maven-jdocbook-plugin:2.1.2 or one of its dependencies could not be resolved: Failed to collect dependencies at org.jboss.maven.plugins:maven-jdocbook-plugin:jar:2.1.2 -> org.jboss:jbossorg-docbook-xslt:jar:1.1.0: Failed to read artifact descriptor for org.jboss:jbossorg-docbook-xslt:jar:1.1.0: Could not transfer artifact org.jboss:jbossorg-docbook-xslt:pom:1.1.0 from/to old jboss.org (http://repository.jboss.org/maven2): Access denied to: http://repository.jboss.org/maven2/org/jboss/jbossorg-docbook-xslt/1.1.0/jbossorg-docbook-xslt-1.1.0.pom , ReasonPhrase:Forbidden. -> [Help 2]
[ERROR] Unknown packaging: jdocbook @ org.jboss:User-Guide-${translation}:1.3.1.Final-SNAPSHOT, /home/pi/mod_cluster/mod_cluster/docs/userguide/pom.xml, line 11, column 14
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/ProjectBuildingException
[ERROR] [Help 2] http://cwiki.apache.org/confluence/display/MAVEN/PluginResolutionException

2.0.0.Alpha1 version is maintained on master branch

I noticed that support for tomcat 8.5 was only added on the master branch. Are the features included on master stable? If so, I was curious if a separate branch and tag could be created for the 2.0.0.Alpha1 version, similar to how the previous 1.3.x was maintained. I would like to leverage this product but wanted a way to ensure my build was reproducible.

mod_cluster doesn't consider parallel deployment while removing context

We use mode_cluster together with a cluster of Tomcat containers. Each container might have multiple versions of the same app deployed (using the parallel deloyment of Tomcat together with automatic undeployment of old versions). Now consider the usecase:

• new version of existing app is deployed to Tomcat
• old app loses all the active sessions
• the old app is undeployed
• mod_cluster removes the corresponding context even though there is still one existing (the new version)

mod_cluster 1.4.1
Tomcat 9.0.21

mod_cluster_slotmem.so failed to load into httpd v2.2.22

I tried to upgrade mod_cluster from v1.2.x to v1.3.1.final on Apache v2.2.22, which is satisfied with the lowest requirements.

Unfortunately, when I restart the apache, I got the error message like:
/opt/app/apache2/modules/mod_cluster/mod_cluster_slotmem.so: undefined symbol: ap_unixd_config

I have searched around for this issue, but no available answers found.

Please kindly refer me to correct solutions. Thanks.

Integrate mod_cluster with embedded Tomcat 8

I want integrate mod_cluster with embedded Tomcat 8. Do you have any advise? I'll be very grateful for your help.

try download, compile and use mod_cluster-1.3.20.Final but not work

trying these commands::

# wget https://github.com/modcluster/mod_cluster/archive/refs/tags/1.3.20.Final.tar.gz
# tar -xf 1.3.20.Final.tar.gz
# cd mod_cluster-1.3.20.Final/native/
# mkdir build
# cd build/
# cmake ../ -G "Unix Makefiles"
# make
# cp modules/*.so /etc/httpd/modules/
# systemctl start httpd
# systemctl start wildfly

but in apache console i see:
[mpm_prefork:notice] [pid 20745] AH00163: Apache/2.4.58 () OpenSSL/1.0.2k-fips mod_cluster/1.3.20.Final configured -- resuming normal operations

then
AH01144: No protocol handler was valid for the URL /rest/api/v1/autenticate
AH01144: No protocol handler was valid for the URL /app.css
AH01144: No protocol handler was valid for the URL /app.js
AH01144: No protocol handler was valid for the URL /..........................

only version that works is:
[mpm_prefork:notice] [pid 26068] AH00163: Apache/2.4.58 () OpenSSL/1.0.2k-fips mod_cluster/1.3.11.Final configured -- resuming normal operations

what am I doing wrong ?

Links at http://mod-cluster.jboss.org/ are broken

Various links at http://mod-cluster.jboss.org/ are broken, they are prepending an extra "/mod_cluster/" to the correct addresses.

For example, the "Documentation" link currently points to: http://mod-cluster.jboss.org/mod_cluster/docs which is broken, but the correct address is: http://mod-cluster.jboss.org/docs

mod_proxy_cluster.c:225:5: warning: implicit declaration of function 'ap_proxy_port_of_scheme' [-Wimplicit-function-declaration]

Hi,

Am facing below warning while compiling mod_proxy_cluster module for apache httpd, with the below warning the mod_proxy_cluster module (mod_proxy_cluster.so) was generated.
the remaining modules advertise, mod_cluster_slotmem and mod_manager able to compile with out any error or warning.

[root@localhost mod_proxy_cluster]# make
/usr/lib64/apr-1/build/libtool --silent --mode=compile gcc -std=gnu99 -pthread -DLINUX -D_REENTRANT -D_GNU_SOURCE -I/opt/apache2/include -I. -I/usr/include/apr-1 -I../include -prefer-pic -c mod_proxy_cluster.c && touch mod_proxy_cluster.slo
mod_proxy_cluster.c: In function 'create_worker':
mod_proxy_cluster.c:291:5: warning: implicit declaration of function 'ap_proxy_port_of_scheme' [-Wimplicit-function-declaration]
if (uri.port && uri.port == ap_proxy_port_of_scheme(uri.scheme)) {
^
/usr/lib64/apr-1/build/libtool --silent --mode=link gcc -std=gnu99 -pthread -o mod_proxy_cluster.la -rpath /opt/apache2/modules -module -avoid-version mod_proxy_cluster.lo
/opt/apache2/build/instdso.sh SH_LIBTOOL='/usr/lib64/apr-1/build/libtool --silent' mod_proxy_cluster.la pwd
/usr/lib64/apr-1/build/libtool --silent --mode=install install mod_proxy_cluster.la /home/mod_cluster-1.3.10.Final/native/mod_proxy_cluster/
libtool: install: warning: remember to run `libtool --finish /opt/apache2/modules'

Apache httpd version
[root@localhost bin]# ./httpd -v
Server version: Apache/2.4.6 (Unix)
Server built: Apr 25 2020 21:44:42

gcc version 4.8.5 20150623 (Red Hat 4.8.5-39) (GCC)

mod_cluster-1.3.10.Final.zip

After integration mod cluster modules (mod_advertise.so, mod_cluster_slotmem.so, mod_manager.so and mod_proxy_cluster.so) with Apache httpd, the mod_proxy_cluster.so module failed to load and getting below error from Apache httpd.
[root@localhost bin]# ./apachectl start
httpd: Syntax error on line 152 of /opt/apache2/conf/httpd.conf: Cannot load modules/mod_proxy_cluster.so into server: /opt/apache2/modules/mod_proxy_cluster.so: undefined symbol: proxy_module

Can you help provide the work around to fix the warning

download link is broken

http://www.jboss.org/mod_cluster is a broken link

Error 526 - Invalid SSL certificate on https://modcluster.io/

Hi,

Since a couple of days, I see this error when trying to access the https://modcluster.io/ website :

Is there another way to see modcluster documentation ?

Florian.

Handling for unavailable nodes

We use Wildfly 10 with modcluster to balance a couple of app server instances. Now, we are looking for a way to tell the load balancer that a node is in trouble and should be removed from the list of balancing members.

I did not find a (documented) way to this in an elegant manner. So, I added a custom LoadMetric that reports errors by throwing an exception (at least this is what the API invites to). However, my exception is swallowed in DynamicLoadBalanceFactorProvider#getLoadBalanceFactor().

The point is: Is there a possibility to invoke a kind of DISABLE-APP event or something similar? If not, a nicer exception handling in DynamicLoadBalanceFactorProvider would be desirable. For example, returning -1 in case of an exception will lead to disablement of the cluster member.

Rather odd thing when mod_cluster is port-mapped

I manually built a 1.3.1 mod_cluster to use against the latest apaches some months ago. Mostly this works fine, but I have a very strange issue.

I run the apache/mod_cluster combo inside a docker.

If I run it and I don't explicitly map the (internal) ports, docker will pick assign some random ones automatically.

CONTAINER ID        IMAGE                COMMAND               CREATED             STATUS              PORTS                                                               NAMES
2bde06b591d9        rt_cluster:latest    "/bin/bash /run.sh"   11 minutes ago      Up 11 minutes       0.0.0.0:49154->80/tcp, 0.0.0.0:49155->6666/tcp, 0.0.0.0:49161->22/tcp   rt-frontend

I can connect to the port 49154, and sticky sessions work correctly, and failover works correctly.

however. If I manually map either of those two ports (-p 80:80), badness happens. If I stop the host that my jsessionId cookie is bound to, instead of failing over, nothing happens on the client. Even more strangely, I get a segfault in the logs:

[Tue Oct 14 15:38:32.236872 2014] [core:notice] [pid 9:tid 140433123825536] AH00052: child pid 69 exit signal Segmentation fault (11)
[Tue Oct 14 15:38:32.243495 2014] [:error] [pid 98:tid 140432948328192] proxy: CLUSTER: (balancer://rtcluster). All workers are in error state for route (155dc85835c7)
[Tue Oct 14 15:38:34.238732 2014] [:notice] [pid 98:tid 140433037821696] Balancer rtcluster changed

I cannot for the life of me guess at how the environment is different when a specific port is mapped through, so this is a query just in case someone has seen this before, or has a good guess.

I'm going to try to re-build 1.3.1 from master and see if it's still an issue.

Apache httpd module releases

Hi,

the documentation talks about downloading a mod_cluster release (https://docs.modcluster.io/#downloads) and using it by unzipping (https://docs.modcluster.io/#install-the-whole-httpd).

The download links here https://www.modcluster.io/downloads/ however only direct to this GitHub repository, which I can only locate source tarballs in.

Am I missing something / should the documentation be updated to reference build instructions instead?
The README in this repository seems to furthermore only reference the JBOSS implementation, not the Apache modules.

Thanks for any pointers!

README possibly out of date

In the README, there are several comments such as

Distribution archives are provided for each Tomcat version.

but I don't see any.

Does this still hold true? If yes, I would suggest clarifying (or even better adding links) , otherwise we should revise this.

mod_cluster crashes apache thread if multiple nodes report the same AJP connect string

We ran into what was ultimately a configuration error in our chef setup but because of how mod_cluster failed, it took us a while to figure out what the problem really was. It presented as random crashes in apache processes with the following two lines in error.log:

[notice] child pid 12758 exit signal Aborted (6), possible coredump in /etc/apache2
apache2: misc/apr_reslist.c:159: reslist_cleanup: Assertion `rl->ntotal == 0' failed.

The ultimate problem was that our wildfly instances were being set up to bind to 0.0.0.0 and then were reporting that IP to mod_cluster. Detail that may be relevant here: We are not using multicast for discovery. We are explicitly setting the IP of the apache servers in the wildfly config like so:

<subsystem xmlns="urn:jboss:domain:modcluster:1.2">
    <mod-cluster-config advertise-socket="modcluster" proxy-list="10.133.248.4:6666,10.133.248.5:6666,10.133.248.6:6666,10.133.248.7:6666" advertise="false" load-balancing-group="${mycluster.modcluster.lbgroup:StdLBGroup}" connector="ajp">
        <dynamic-load-provider>
            <load-metric type="busyness"/>
        </dynamic-load-provider>
    </mod-cluster-config>
</subsystem>

Anyway, the end result is that the mod_cluster-manager page was displaying this (node details removed for brevity):

Now, clearly this configuration is wrong. The only reason it even worked in the first place to serve up any content was that wildfly is currently running on the same host as apache so sending requests to 0.0.0.0 actually works. If this was not the case, requests would have thrown 404s or 503s or something and the problem would have been obvious. But the setup worked well enough that we thought maybe it was some kind of mismatch between apache and mod_cluster since we were using an older version of mod_cluster initially. So we spent time trying to chase down possible bugs or issues in our installation process.

Our current theory is that the crashes happened when mod_cluster decided to switch a user to a different node for load balancing reasons and then either the duplicated connect string or the fact that when it tried to talk to a different node, it got the same one on 0.0.0.0 caused the crash.

I'm not sure what would constitute a fix for this. One simple solution might just be to have mod_cluster complain loudly if two nodes register themselves with the same connect string. Is there any valid reason for this to ever happen? One line in error.log would have saved us hours of debugging.

Native Sources Missing in Master Branch

Documentation @ http://docs.jboss.org/mod_cluster/1.3.0/html/native.building.html#d0e770 says to build from git master branch. Master branch does not contain the native source directories to build from as described in documentation. I was able to find the native sources directory in the 1.3.x branch and build from there. Not sure if the repo is functioning as expected and the documentation needs to be updated, or vice versa. Please update either one as required, thank you!

issue related to starting apache after configuring for mod_cluster

I am configuring mod_cluster, using apache, Jboss EAP 6.3. Extracted mod_cluster-1.2.6.Final-linux2-x64-so.tar and copied :mod_advertise.so, mod_manager.so, mod_slotmem.so, mod_proxy_cluster.so into /etc/httpd/modules directory.
Made following changed to httpd.conf file:
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_ajp_module modules/mod_proxy_ajp.so
#LoadModule slotmem_module modules/mod_slotmem.so
LoadModule manager_module modules/mod_manager.so
#LoadModule proxy_cluster_module modules/mod_proxy_cluster.so

ManagerBalancerName mycluster

Listen 127.0.0.1:1001
<VirtualHost 127.0.0.1:1001>

Order deny,allow
Allow from all

KeepAliveTimeout 300
MaxKeepAliveRequests 0
#AdvertiseFrequency 5
#ServerAdvertise off
EnableMCPMReceive On

<Location /mod_cluster-manager>
SetHandler mod_cluster-manager
Order deny,allow
Allow from all
Require all granted

While running the apache server, I am getting following error. Apache server is not coming up:
httpd: Syntax error on line 357 of /etc/httpd/conf/httpd.conf: Cannot load modules/mod_manager.so into server: /etc/httpd/modules/mod_manager.so: undefined symbol: ap_my_generation
What is the solution to fix above issu

[1.3] Push SCM tag automatically during release

ERROR MODCLUSTER000042

ERROR [org.jboss.modcluster] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) MODCLUSTER000042: Error null sending INFO command to X.X.X.X/X.X.X.X:6666, configuration will be reset: null
This error is why? many thanks!

mod_cluster_slotmem.so failed to load into httpd 2.4.6-19 and 2.4.6-30

I tried to download every release available here even the 1.3.1 Beta2 unsuccessfully.
I tried to compile them manually or just downloading the binary "so" but everytime my httpd failed to start with this error message:

Can't locate API module structure `slotmem_module' in file /etc/httpd/modules/mod_cluster_slotmem.so: /etc/httpd/modules/mod_cluster_slotmem.so: undefined symbol: slotmem_module

others modules load correctly except for this one!
Is there a solution/workaround
Thanks a lot.

Docs link to Advertize.java is broken

On the docs page https://docs.modcluster.io/#_it_is_not_working_what_should_i_do

Check that Advertise message are received on the cluster node. A small
 Java utility could be used to check Advertise. It is in the 
mod_cluster repository and can be compiled using javac. The 
output should be something like:

The text small Java utility links to https://github.com/modcluster/mod_cluster/blob/master/test/java/Advertize.java which is 404.

1. Did Advertize.java move? I cannot find it.
2. Was it replaced by a new test fixture?

The doc needs to be updated to reflect these changes.

No sign version

Hi guys,

How to check the integrity of your tar ?
Is there a way to sign with something like pgp ?

Regards,

Can not Forwarding from www to non-www version while using SSL with modcluster

I configured httpd with modcluster and I want to redirect www to non-www version while using SSL but it did not work.
My current configuration work as follow
http://www.qoosi.com => https://qoosi.com ( not work )
http://qoosi.com => https://qoosi.com ( work)
https://www.qoosi.com => https://qoosi.com ( not work)
Here is the configuration for modcluster and a virtual host listen on port 80 to forward from http to https:

<IfModule manager_module>
    Listen 192.168.123.9:6666
    ManagerBalancerName wfycluster
    <VirtualHost 192.168.123.9:6666>
        AllowDisplay On
        ServerAdvertise On
        KeepAliveTimeout 300
        MaxKeepAliveRequests 0
        AdvertiseFrequency 5
        ServerAdvertise On 192.168.123.9:6666
        AdvertiseGroup 224.0.1.105:23364
        EnableMCPMReceive On
        ProxyPassMatch ^/sas/ !
        ProxyPass / balancer://wfycluster stickysession=JSESSIONID|jsessionid nofailover=on
        ProxyPassReverse / balancer://wyfcluster
        ProxyPreserveHost on
        <Location />
            Require all granted
        </Location>
        <Location /wfycluster>
            SetHandler mod_cluster-manager
            Require all granted
        </Location>
    </VirtualHost>
</IfModule>
<VirtualHost *:80>
<VirtualHost *:80>
    RewriteEngine On
    RewriteCond %{HTTP_HOST} ^www\.(.*)$ [NC]
    RewriteRule ^(.*)$ https://%1/$1 [R=301,L]
</VirtualHost>

Here is the configuration for SSL virtual-host(ssl.conf):

Listen 443 https
SSLPassPhraseDialog exec:/usr/libexec/httpd-ssl-pass-dialog
SSLSessionCache         shmcb:/run/httpd/sslcache(512000)
SSLSessionCacheTimeout  300
SSLCryptoDevice builtin
<VirtualHost _default_:443>
#DocumentRoot "/var/www/html"
ServerName qoosi.com:443
ServerAlias wwww.qoosi.com:443
RewriteEngine On
RewriteCond %{HTTPS} off
RewriteCond %{ENV:HTTPS} off
RewriteRule ^ https://%{HTTP_HOST}%{REQUEST_URI} [L,R=301,NE]
RewriteCond %{HTTP_HOST} ^www\.(.+)$ [NC]
RewriteRule ^ https://%1%{REQUEST_URI} [L,R=301,NE]
       ProxyPassMatch ^/sas/ !
       ProxyPass / balancer://wfycluster stickysession=JSESSIONID|jsessionid nofailover=on
       ProxyPassReverse / balancer://wyfcluster
       ProxyPreserveHost on
#static asset  
alias /sas  /var/www/html/sas
ErrorLog logs/ssl_error_log
TransferLog logs/ssl_access_log
LogLevel warn
SSLEngine on
SSLProtocol all -SSLv2
SSLCipherSuite HIGH:MEDIUM:!aNULL:!MD5:!SEED:!IDEA
SSLCertificateFile /etc/httpd/ssl/qoosi_com.crt
SSLCertificateKeyFile /etc/httpd/ssl/qoosi_com.key
#SSLCertificateChainFile /etc/pki/tls/certs/server-chain.crt
SSLCACertificateFile /etc/httpd/ssl/qoosi_com.ca-bundle
#SSLVerifyClient require
#SSLVerifyDepth  10
 
<Files ~ "\.(cgi|shtml|phtml|php3?)$">
    SSLOptions +StdEnvVars
</Files>
<Directory "/var/www/cgi-bin">
    SSLOptions +StdEnvVars
</Directory>
BrowserMatch "MSIE [2-5]" \
         nokeepalive ssl-unclean-shutdown \
         downgrade-1.0 force-response-1.0
 
</VirtualHost>