Steps that I followed to install KINSOLVER in a linux-box 2.6.32-358.6.2.el6.x86_64
- Get the KINSOLVER C++ source code
- Created folder ~/kin_c_source
- Downloaded Makefile, kin.cpp, kin.h
- Compiled executable via command-line: make
- Make folder ~/kin
- Copy executable kin_c to ~/kin/
- The backend simulator (executable) is then located at ~/kin/kin_c
- Get the Java KINSOLVER source code
- Created folder kin_java_source
- Downloaded Java files (Dataset.java, KinFile.java, …, and )
- Update PlotServlet.java to make sure that the URL is updated accordingly (e.g., change from genetics.uga.edu to your own server)
- Update the script to compile (compile.sh)
- The jar file for servlets may need to be specifies, such as using the one in apache tomcat. Example: /opt/apache-tomcat-7.0.40/lib/servlet-api.jar
- Make sure Gnuplot works
- For example, admin had it installed at: /usr/bin/gnuplot
- Setup java-servlets
- Inside folder ‘jsdk2.1’, update variables inside of file: kin_environment.properties
- Start the servlet-server by running ./startserver
- Verify that the servlets are running by going to http://odyssey.bioinformatics.uga.edu:8090/kin/servlet/Setup , which is a servlet that keeps track of the options needed to run KINSOLVER
- In the case that the browser takes a while to show anything and times out, then likely the port (i.e., 8090) is not accessible either in the network (such as inside campus computers) or the linux-box has the port closed (system-administrator needs to open the port)
- Go to servlet with configuration options, save changes, and click on “create web page”
- Test via uploading an input file
- Troubleshoot until it works
- For example, I changed ‘kinnew’ to ‘kin’ in order to have simpler URLs
- Reminder: changes on java-servlets code requires stopping the server and starting it up again to be visible
- Inside folder ~/jsdk2.1
- Start server: ./startserver &
- Stop server: ./stopserver
- In case of error messages when stopping it, just ignore them.