Optimize space consumption during testing jobs

Currently, jobs like the following:
https://qa2.nuxeo.org/jenkins/view/Depl/job/Deploy/job/IT-release-on-demand-tests-cap-tomcat-nx810/

echo $NX_CLID | sed 's/--/\n/g' > tomcat/nxserver/data/instance.clid
cp -R tomcat tomcat-cmis
cp -R tomcat tomcat-funkload
cp -R tomcat tomcat-webdriver
mv tomcat tomcat-wizard

are duplicating the servers for running several tests.
The problem is it takes 4 times the space needed because the 4 test suites run sequentially thus don't need the datas until they begin.

The same applies to the following jobs through scripting;
https://qa2.nuxeo.org/jenkins/view/Private/job/Private/job/HFBuilding/job/IT-HF-PKG-tests-standalone-tomcat-nx810/configure

in:
https://github.com/nuxeo/integration-scripts-priv/blob/master/libs/misc.sh

    rm -rf $L_WORKSPACE/tomcat-*
    cp -R $L_NX_INSTANCE $L_WORKSPACE/tomcat-funkload
    cp -R $L_NX_INSTANCE $L_WORKSPACE/tomcat-cmis
    cp -R $L_NX_INSTANCE $L_WORKSPACE/tomcat-wizard
    cp -R $L_NX_INSTANCE $L_WORKSPACE/tomcat-webdriver

This could be improved by preparing the data only before each test suite. This would divide the disk space needed by roughly 4 times in each job concerned.

Database tests for example run 2 test suites so twice the space needed.

The situation is even worse when 2 HF releases are in progress the same day.