Upgrade to TeamForge 8.1 - Database and SCM on separate servers

In this procedure, we'll assume that you are upgrading on the same server where your existing TeamForge 8.0 site is running.

Important: TeamForge 8.1 supports RHEL/CentOS 7.1. However, it also supports RHEL/CentOS 6.6 if you choose to upgrade on the same server that runs RHEL/CentOS 6.6.
In this option, the following services run on the application server (we call this my.app.host). The following service runs on the database server (we call this my.db.host). The following services run of the SCM server (we call this my.scmandgit.host).
Note: For the ETL service to run as expected in a distributed TeamForge installation, all servers must have the same time zone.

Log on to the server as root user always.

Do the following on the application server - my.app.host

  1. Uninstall the TeamForge CLI add-on.
    • cd /opt/collabnet/teamforge/add-ons/teamforge_cli_server
    • ./install --uninstall
    • yum erase teamforge_cli_server
  2. Stop TeamForge.
    • /etc/init.d/collabnet stop all
  3. No backup is required for same hardware upgrades. However, you can create a backup as a precaution. See Back up and restore CollabNet TeamForge data to learn more about backing up TeamForge database and file system.
  4. Upgrade the operating system packages.
    • yum upgrade
  5. Create a file, /etc/security/limits.d/teamforge.conf, and make sure it has the following "nproc" and "nofile" settings.
    • vim /etc/security/limits.d/teamforge.conf
    • sf-admin soft nproc 10240
    • sf-admin hard nproc 10240
    • sf-admin hard nofile 65536
    • sf-admin soft nofile 65536
    • Save the file and reboot the server.
  6. If the TeamForge server has SELinux enabled, run it in 'Permissive' mode temporarily while installing or upgrading TeamForge.
    1. Verify if SELinux is running in enforcing mode.
      • getenforce
    2. If the output of the getenforce command is "Permissive", continue with the next step. If not, run the following command to bring it to 'Permissive' mode.
      • setenforce 0
    See Set up SELinux to have TeamForge run in SELinux mode after completing the installation or upgrade.
  7. Important: Do this only if you are upgrading TeamForge on RHEL/CentOS 6.6.
    Delete the python-crypto package.
    • yum erase python-crypto
  8. Configure your TeamForge 8.1 installation repository. See TeamForge installation repository configuration for Red Hat/CentOS.
  9. Install the following application packages.
    1. TeamForge: To install the TeamForge application packages run the following command:
      • yum install teamforge collabnet-binary
    2. Black Duck Code Sight: To install the Black Duck Code Sight packages run the following command:
      • yum install teamforge-codesearch
  10. In the site-options.conf file, make sure you do the following.
    Note: Back up your site-options.conf file before making any changes.
    • vi /opt/collabnet/teamforge-installer/8.1.0.1/conf/site-options.conf
    1. Update the host name and domain name, if required.
      HOST_localhost=app etl indexer
      HOST_my.db.domain.com=database datamart 
      HOST_my.scmandgit.domain.com=subversion cvs gerrit
      DOMAIN_localhost=my.app.domain.com
      Add 'codesearch' to the HOST_localhost token if you are installing Black Duck Code Sight.
      HOST_localhost=app etl indexer codesearch
      Add 'binary' to the HOST_localhost token if you are installing Nexus.
      HOST_localhost=app etl indexer codesearch binary
    2. Turn on the SSL for your site by editing the relevant variables in the site-options.conf file. To generate the SSL certificates, see Generate SSL certificates.
      • SSL=on
      • SSL_CERT_FILE
      • SSL_KEY_FILE
      • SSL_CHAIN_FILE
      Note: The SSL_CERT_FILE and SSL_KEY_FILE tokens need an absolute path. The SSL_CHAIN_FILE token is optional.
    3. If your site is running in SSL mode (SSL=on), add the following java runtime property to the JBOSS_JAVA_OPTS token.
      JBOSS_JAVA_OPTS=-Xms1536m -Xmx1536m -XX:MaxPermSize=512m -server -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/tmp -verbose:gc -XX:+
      PrintGCTimeStamps -XX:+PrintGCDetails -Djsse.enableSNIExtension=false -Dsun.rmi.dgc.client.gcInterval=600000 -Dsun.rmi.dgc.server.gcInterval=600000
      
    4. If the token REQUIRE_PASSWORD_SECURITY is enabled, then set a value for the token, PASSWORD_CONTROL_EFFECTIVE_DATE.
      CAUTION:
      The Password Control Kit (PCK) disables, deletes or expires user accounts that don't meet the password security requirements starting from the date set for the PASSWORD_CONTROL_EFFECTIVE_DATE token. If a date is not set, the PCK disables, deletes or expires user accounts immediately. See PASSWORD_CONTROL_EFFECTIVE_DATE for more information.
    5. TeamForge 7.1 and later support automatic password creation. See AUTO_DATA for more information.
    6. Password obfuscation

      The password obfuscation is enabled by default. As a result, all password-related tokens are encrypted in all the TeamForge configuration files.

      Restriction: The password-related tokens cannot contain the following characters in the site-options.conf file: $<>/\'"`
      • To disable password obfuscation, set OBFUSCATION_ENABLED=false.
      • To configure the obfuscation key, set OBFUSCATION_KEY=<Any AlphaNumeric value with length >= 8 bytes>. The default value of OBFUSCATION_KEY token is XSJt43wN.
    7. Ensure to set the token DEDICATED_INSTALL=true.
    8. Include the list of non-expiring TeamForge user accounts (password never expires).
      USERS_WITH_NO_EXPIRY_PASSWORD=admin,nobody,system,scmviewer,scmadmin
    9. Make sure that the following tokens have a value if ETL is enabled.
      SOAP_ANONYMOUS_SHARED_SECRET
      ETL_SOAP_SHARED_SECRET
    10. Configure Black Duck Code Sight tokens. See Black Duck Code Sight site-option tokens.
    11. To enable the history protection feature of TeamForge Git integration, set the GERRIT_FORCE_HISTORY_PROTECTION=true. For more information, see GERRIT_FORCE_HISTORY_PROTECTION.
    12. If you have LDAP set up for external authentication, you must set the “REQUIRE_USER_PASSWORD_CHANGE” site options token to false.
    13. Ensure to set the token SELINUX_SETUP=true.
    14. Important: This step is required if you want Git notification emails.
      Update the JAMES_ACCEPTED_RELAYS site-options token with the Git server's IP address. See JAMES_ACCEPTED_RELAYS for more information.
      JAMES_ACCEPTED_RELAYS=127.0.0.1,{__CEE_DOMAIN__},<The IP address of the Git server>
    15. Save the site-options.conf file.
  11. As JDK is upgraded to 1.8.0_45, move the existing runtime directory before create-runtime.
    • cd /opt/collabnet/teamforge/
    • mv runtime/ runtime_old
  12. Recreate the runtime environment.
    • cd /opt/collabnet/teamforge-installer/8.1.0.1
    • ./install.sh -r -I -V
  13. Important: Do this if you have updated the JAMES_ACCEPTED_RELAYS token with the Git server's IP address.
    Edit the /opt/collabnet/teamforge/runtime/james/apps/james/SAR-INF/config.xml file and comment out the <authorizedAddresses> node. For example:
    <!--  <authorizedAddresses>127.0.0.0/8</authorizedAddresses>   -->
  14. Update the file permissions on your site's data.
    • /opt/collabnet/teamforge/runtime/scripts/fix_data_permissions.sh
    Note: This process can take a long time for sites with a lot of data.

Do the following on the database server - my.db.host

  1. Run the following commands to upgrade Red Hat/CentOS to the latest version.
    1. Remove the neon-devel package if you are upgrading from TeamForge 6.2.
      • yum erase neon-devel -y
      • yum erase subversion-devel -y
    2. Upgrade the operating system packages.
      • yum upgrade
  2. If the TeamForge server has SELinux enabled, disable it temporarily while installing or upgrading TeamForge.
    1. Verify if SELinux is running in enforcing mode.
      • getenforce
    2. If the output of the getenforce command is either "Disabled" or "Permissive", SELinux is already disabled.
    3. If not disabled, run the following command to disable SELinux.
      • setenforce 0
  3. Configure your TeamForge 8.1 installation repository. See TeamForge installation repository configuration for Red Hat/CentOS.
  4. If the PostgreSQL database is running locally, stop the PostgreSQL service.
    • /etc/init.d/postgresql-9.3 stop
  5. Install the TeamForge database packages.
    • yum install teamforge-database
  6. Copy the site-options.conf file from my.app.host and modify the token settings.
    Note: If you choose to use the old site-options.conf file, don’t forget to copy the AUTO_DATA token from the application server.
    1. Set the following tokens.
      HOST_my.db.host=database datamart
      Note: 'HOST_my.db.host' is just an example. As you are installing database on a separate server, do not use 'HOST_localhost'. Use 'HOST_<valid host name>' instead.
      DOMAIN_my.db.host=my.db.domain.com
      HOST_my.app.domain.com=app etl indexer
      HOST_my.scmandgit.domain.com=subversion cvs gerrit
    2. Save the site-options.conf file.
  7. As JDK is upgraded to 1.8.0_45, move the existing runtime directory before create-runtime.
    • cd /opt/collabnet/teamforge/
    • mv runtime/ runtime_old
  8. Recreate the runtime environment.
    • cd /opt/collabnet/teamforge-installer/8.1.0.1
    • ./install.sh -r -I -V

Do this on the application server - my.app.host

  1. Convert your site data to work with TeamForge 8.1.
    Tip: Before you kick off the data migration, do the following:
    • Use the /etc/init.d/collabnet status command to make sure the JBoss, Tomcat and Tomcatcs services are stopped.
    • If any of the trackers in your project already contains a field with the name team, conflict may arise while running the migrate.py script. This is because of the inclusion of the Team feature in TeamForge 8.0. To prevent this conflict, follow the steps given in FAQ on troubleshooting.
    • /opt/collabnet/teamforge/runtime/scripts/migrate.py
    The migrate.py script locates the existing site data and modifies it as needed.

    This includes configuration data for LDAP and the James mail server. Any modifications that you have applied to these components on your old site are reproduced on your upgraded TeamForge 8.1 site.

  2. Start TeamForge.
    • /etc/init.d/collabnet start
  3. If you have CVS integrations, synchronize permissions post upgrade. See, Synchronize TeamForge source control integrations.
  4. Run the TeamForge post installation script. For more information, see post-install.py.
    • /opt/collabnet/teamforge/runtime/scripts/post-install.py
    Note: In case you face any GIT-related issues while running the post-install.py script, see Post install fails for GIT. What should I do?.
  5. Important: This step is required if and only if you have reporting enabled and are upgrading to TeamForge 8.1.
    If you have reporting enabled, it is more likely that you would have run initial load jobs and set up incremental load jobs on your site. See etl-client.py for more information on initial and incremental load jobs. TeamForge 8.1 and later have custom-defined tracker fields included in the datamart.
    1. Run the following commands assuming you have tracker incremental load job set up on your site and that you just want to have your datamart loaded with prospective updates to custom-defined tracker fields. Your incremental load job would fail if the following commands are not run.
      • cd /opt/collabnet/teamforge/runtime/scripts
      • ./etl-client.py –a
      • ./etl-client.py –r LoadFlexFields
      Going forward, the tracker incremental load job loads all prospective changes to custom-defined tracker fields to your datamart.
    2. On the other hand, you can bootstrap your datamart and manually run the initial load jobs again, so that your datamart is loaded with data from historical custom-defined tracker fields, while at the same time is ready for prospective updates as well.
      CAUTION:
      However, keep in mind you are bootstrapping your datamart (reporting DB) and you may have to run the tracker initial load job anyway, which may take a while to complete depending on your volume of data during which your reporting would not be available. You may also lose all the data related to UserLoggedInActivity. Run the following commands if you understand the situation at hand.
      • /etc/init.d/collabnet stop
      • /opt/collabnet/teamforge/runtime/scripts/bootstrap-reporting-data.py
      • /etc/init.d/collabnet start
      • /opt/collabnet/teamforge/runtime/scripts/etl-client.py –r SCMCommitActivityJob
      • /opt/collabnet/teamforge/runtime/scripts/etl-client.py –r TrackerInitialJob
  6. Install the Black Duck Code Sight license on the server where Black Duck Code Sight is installed. For more information see these instructions.
  7. Run the svn_cache.sh script.
    • cd /opt/collabnet/teamforge/runtime/scripts/codesearch/
    • ./svn_cache.sh <Repository Base URL Path of the SCM Integration Server>

    Provide a repository base URL path of the SCM integration server, for example, "http://myint.box.net/svn/repos", where myint.box is the server with the SCM integration server.

    In addition, if you add a new integration server at some point later, you must run this svn_cache.sh script, (after creating the new integration server), on the TeamForge application server.

Do the following on the SCM server - my.scmandgit.host

If your TeamForge setup includes source control running on its own server, you'll have to upgrade that server as well as the main TeamForge application server.

  1. Stop TeamForge.
    • /etc/init.d/httpd stop
    • /etc/init.d/collabnet stop tomcat
  2. No backup is required for same hardware upgrades. However, you can create a backup as a precaution. See Back up and restore CollabNet TeamForge data to learn more about backing up TeamForge database and file system.
  3. Run the following commands to upgrade Red Hat/CentOS to the latest version.
    1. Remove the neon-devel package if you are upgrading from TeamForge 6.2.
      • yum erase neon-devel -y
      • yum erase subversion-devel -y
    2. Upgrade the operating system packages.
      • yum upgrade
  4. If the TeamForge server has SELinux enabled, disable it temporarily while installing or upgrading TeamForge.
    1. Verify if SELinux is running in enforcing mode.
      • getenforce
    2. If the output of the getenforce command is either "Disabled" or "Permissive", SELinux is already disabled.
    3. If not disabled, run the following command to disable SELinux.
      • setenforce 0
  5. Important: Do this only if you are upgrading TeamForge on RHEL/CentOS 6.6.
    Delete the python-crypto package.
    • yum erase python-crypto
  6. Configure your TeamForge 8.1 installation repository. See TeamForge installation repository configuration for Red Hat/CentOS.
  7. If Git is enabled and the PostgreSQL database is running locally:
    1. Stop the PostgreSQL service.
      • /etc/init.d/postgresql-9.2 stop
  8. Install the following application packages.
    1. Install the source code component.
      • yum install teamforge-scm
    2. To install the Git packages, run the following command.
      • yum install teamforge-git
  9. Copy the master site-options.conf file from my.app.host and modify the host token settings in the site-options.conf file.
    Note: If you choose to use the old site-options.conf file, don’t forget to copy the AUTO_DATA token from the application server.
    HOST_my.scmandgit.host=subversion cvs gerrit
    DOMAIN_my.scmandgit.host=my.scmandgit.domain.com
    HOST_my.app.domain.com=app etl indexer codesearch
    HOST_my.db.domain.com=database datamart
  10. Save the site-options.conffile.
  11. As JDK is upgraded to 1.8.0_45, move the existing runtime directory before create-runtime.
    • cd /opt/collabnet/teamforge/
    • mv runtime/ runtime_old
  12. Recreate the runtime environment.
    • cd /opt/collabnet/teamforge-installer/8.1.0.1
    • ./install.sh -r -I -V
  13. Restart Apache.
    • RHEL/CentOS 6.x: systemctl reload httpd
    • RHEL/CentOS 7.x: /etc/init.d/httpd reload
    Important: Make sure you restart Apache whenever you do a runtime recreation on the SCM server.
  14. Update the file permissions on your site's data.
    • /opt/collabnet/teamforge/runtime/scripts/fix_data_permissions.sh
    Note: This process can take a long time for sites with a lot of data.
  15. Start TeamForge.
    • /etc/init.d/collabnet start
  16. Run the TeamForge post installation script. For more information, see post-install.py.
    • /opt/collabnet/teamforge/runtime/scripts/post-install.py
    Note: In case you face any GIT-related issues while running the post-install.py script, see Post install fails for GIT. What should I do?.
  17. Important: This step is required if you want Git notification emails.
    Edit the /opt/collabnet/gerrit/etc/gerrit.config file and update the 'smtpServer' property with the TeamForge application server's host name. For example:
    [sendemail]
                  smtpServer = <TeamForge application server host name>
    1. Restart gerrit.
      • /etc/init.d/collabnet stop gerrit
      • /etc/init.d/collabnet start gerrit

Do this on the application server - my.app.host

  1. Apply the finishing touches and make sure everything is running smoothly after upgrading to TeamForge 8.1.
    Important: Do not delete the teamforge-installer/8.1.0.1 directory. You will need it for future maintenance and upgrades.
    1. Log into your site as the administrator.
    2. If your site has custom branding, verify that your branding changes still work as intended. See Customize anything on your site.
    3. Let your site's users know they've been upgraded. See Communicate with all site users at once.

Do this on the database server - my.db.host

  1. Remove the backup files, if any, after the TeamForge site is up and running as expected. Remove the repository and the file system backup from the /tmp/backup_dir directory.
Note: After upgrade, it takes some time for the publishing repositories to get created for projects imported from other TeamForge sites.