Skip to main content
Version: 10.2.8

Upgrade from 10.1.x to 10.2.8.x

The guide describes upgrading to the 10.2.8.х versions of the Work.AI platform. Note that this operation uses the installer, so the procedure is similar to the installation.

Before you start​

The 10.2.8.х non-HA topology requires five servers:

  • MASTER
  • AGENT
  • RPA
  • BI
  • DB

Note the following changes in server roles:

  • The separate INT, APP, and BEP Master servers no longer exist. Instead, WorkFusion has introduced a single Master server. It contains all services that previously ran on the three servers. For the single-point installation, MASTER becomes the server to run the setup on.
  • BEP Agent servers are renamed to AGENT servers.

The upgrade starts with preparing the environment to bring the topology to the appropriate state.

Before the upgrade, ensure that:

  • You created the additional DNS record for Keycloak and the MinIO UI. See DNS names.
  • You allowed additional ports between servers For ZooKeeper and ETCD. See Ports.
  • You added <MSSQL_PM_USER> to SQLAgentUserRole. See Database.
  • You granted view server state permission to <MSSQL_PM_USER>. See Database.
  • No Business Process (BP) or Manual Task is running.
  • You prepared fleets for migration.
  • You backed up the following components:
    • The /opt/workfusion/workfusion and /opt/workfusion/workspace directories
    • Vault properties for all components
    • ZooKeeper settings changed since the previous installation
    • Any system files changed since the previous installation
  • You enabled the SSH access to all Linux hosts.
  • You verified that the SSO or LDAP authentication you enabled in the existing environment is working.
  • You stopped APP and BEP Master servers.
  • You specified the billing_environment_id and billing_customer_id parameters for the Pendo setup identification.
  • You installed the packages on the Superset server. You also mounted the vds-data directory on the Superset server in case you use an external NFS server. For details, see Prepare for installation | Perform additional setup for Superset server.

Back up previous installation​

You can back up the existing INSTALL_DIR and restore the platform to the previous version after the upgrade if something goes wrong. Note that the backup files occupy the same disk space as INSTALL_DIR. Before performing the procedure, make sure that there is enough space on the filesystem.

To roll back to the previous state, back up the following components:

  • MSSQL database. You can do it according to your standard backup process.

  • The INSTALL_DIR directory. You can do it during the upgrade.

  • Shared directories. If you use any existing NAS, back up the following shared directories as per your standard process:

    • /shared and /vds-data for the HA mode
    • /vds-data for the non-HA mode

Save fleet names for Bot Units​

To save the currently used fleet names for Bot Units, do the following:

  1. Log in to Control Tower and click Bot Manager.

  2. In the Bot Manager window, click your RPA server to expand the fleets.

  3. Copy the value into the Fleet column → Actions → Edit Fleet.

  4. In the Edit Fleet window, click Create new, and paste the values copied in step 3.

  5. Repeat steps 3-4 for all fleets.

Check running Business Processes​

Before the upgrade, ensure that no Business Processes are running:

  1. Go to INSTALL_DIR with the installer directory created during the installation of 10.1.x, for example: /opt/workfusion:

    $ cd INSTALL_DIR

  2. Check the output of the following command:

    ./zookeeper/bin/zkCli.sh ls /control-tower/execution/active-instances | egrep -qo '^\[[0-9a-f-]+' && echo 'There are running business processes. Please stop them first to continue upgrade.' || echo 'There are no running business processes. You are OK to continue update procedure.'

Continue the update procedure if the command displays the following message: "There are no running business processes. You are OK to continue update procedure." Otherwise, stop the running processes first.

Stop services​

  1. Stop unnecessary OCR servers.

    Work.AI v10.2.8.x doesn't use dedicated OCR servers, so that you can stop them. For more details, see the Hardware and OS section. If you plan to retain options for rollback, don't stop the OCR server.

    note

    If no rollback is necessary and you want to use OCR servers as AGENT servers, add the OCR hostnames as agent_hostnames in hosts.yml.

  2. Stop the platform services on the rest of the servers:

    Connect to the server with the previous installation package via SSH and run the following commands as <LINUX_INSTALLATION_USER>:

    $ cd PACKAGE_DIR
    $ export ANSIBLE_PRIVATE_KEY_FILE="<PATH_TO_SSH_KEY>"   # Path to SSH key for connection to all Linux servers as WFUSER.
    $ export ANSIBLE_VAULT_PASS="<ANSIBLE_VAULT_PASSWORD>"   # Password for decryption of config.yml.
    $ ./install.sh stop_services full

Prepare application user​

Versions beyond 10.1.6.1 require APPLICATION_USER (WFUSER) to install and run the Product.

To perform the upgrade, provide the following permissions to APPLICATION_USER:

  1. Grant the sudo permissions.

  2. Provide SSH access via the SSH key to all Linux servers with the installation package.

    If you upgrade the system using the multi-point approach, skip this step.

  3. Provide APPLICATION_USER with the owner permissions for PACKAGE_DIR. This directory contains the installer extracted during the initial installation of 10.1.x, for example: /opt/workfusion/wf_installer.

    $ sudo chown -R APPLICATION_USER PACKAGE_DIR

Prepare environment​

note

When updating the DR Failover environment, skip the Prepare environment section and run steps 1-3 from the Deployment section.

Prerequisites​

Before upgrading the platform, ensure that your environment meets the following requirements. Failing to do so results in an extended time of the upgrade and can cause errors:

Prepare certificates​

The server.crt certificate must mandatorily include the end certificate and all intermediate ones if they exist. Ensure that your certificates comply with the requirements.

Then, on the Master server, add ca.crt, server.crt, and server.key to the PACKAGE_DIR/certificates/ directory.

You can skip the step if you don't have those certificates. In this case, self-signed certificates get generated automatically during the installation.

$ ls -la PACKAGE_DIR/certificates/
total 24
drwxr-xr-x.  3 ec2-user ec2-user   78 Sep 24 14:47 .
drwxr-xr-x. 12 ec2-user ec2-user 4096 Sep 27 17:20 ..
-rw-r--r--.  1 ec2-user ec2-user 2984 Sep 24 14:42 ca.crt
-rw-r--r--.  1 ec2-user ec2-user 4835 Sep 24 14:42 server.crt
-rw-r--r--.  1 ec2-user ec2-user 1674 Sep 24 14:42 server.key

Verify prerequisites​

After you prepare the environment for the upgrade, you can use the dedicated checking script for verification. For more information, read the Check readiness for installation instruction.

Download and extract installer​

To download and extract the installer, connect to the server with the installation package from the previous version via SSH and run the following commands as APPLICATION_USER:

  1. Go to INSTALL_DIR that contains the installer directory created during the installation of 10.1.x, for example: /opt/workfusion:

    $ cd INSTALL_DIR

  2. Create the wf_installer_new directory for the new version of the installer. Ensure that the directory is writable for APPLICATION_USER.

    note

    Use only the wf_installer_new name for the directory to store the new installation package. Otherwise, the backup doesn't start.

    $ mkdir wf_installer_new

  3. Download the v10.2.8.x installer to INSTALL_DIR/wf_installer_new:

    $ curl -o '<10.2.8.x-installer-archive>' '<link-to-installer-10.2.8.x>'

    If the Internet is unavailable on the server, download the package to the PACKAGE_DIR directory using any convenient method.

  4. Extract the installer:

    $ tar xzf <10.2.8.x-installer-archive> --strip 1

  5. Copy configuration files for the current environment from INSTALL_DIR/wf_installer to INSTALL_DIR/wf_installer_new:

    $ cp -r INSTALL_DIR/wf_installer/{license.properties,certificates,.agreement,rpa.yml} INSTALL_DIR/wf_installer_new/

  6. Generate new configuration files from the existing ones:

    $ export ANSIBLE_VAULT_PASS="<ansible_vault_password>"   # Password for decryption of config.yml.

    $ cd INSTALL_DIR/wf_installer_new
    $ ./install.sh hosts_yml  generate -e @INSTALL_DIR/wf_installer/hosts.yml
    $ ./install.sh config_yml generate -e @INSTALL_DIR/wf_installer/config.yml

    After executing ./install.sh hosts_yml generate, in hosts.yml, specify the correct value for keycloak_lb_hostname.

    $ ./install.sh ports_yml  generate -e @INSTALL_DIR/wf_installer/ports.yml
    $ ./install.sh rpa_yml    generate -e @INSTALL_DIR/wf_installer/config.yml
    note

    The new mssql_keycloak_user is introduced with the default password in config.yml. Change it to the password of mssql_keycloak_user you created when preparing the database.

    Fill empty passwords in config.yml.

    $ ./install.sh passwords generate
    note

    Alternatively, you can set the passwords manually. In that case, remember to use only Latin letters and allowed symbols as described in the Specify installation parameters guide.

    info

    In 10.2.8.x versions, Work.AI supports Active Directory logins. If you want to switch to MS SQL AD logins, in config.yml, change the corresponding mssql logins to the logins and password you created when preparing the database.

    note

    BI workbooks for Tableau are not published by default during the installation process. If you need them published during the installation, go to config.yml and set the publish_bi_dashboards variable to True. Another option to publish BI workbooks is to run the following command at any time after the installation is completed.

    $ ./install.sh install bi_workbooks -e publish_bi_dashboards=true

  7. Encrypt config.yml:

    $ ./install.sh encrypt config

  8. The server.crt certificate must mandatorily include the end certificate and all intermediate ones if they exist. If your certificates don't comply with the requirements, prepare new ones according to the Certificates guide.

  9. Generate required internal certificates:

    $ ./install.sh certs generate
If you upgrade using the multi-point approach
  1. On the Master server, generate configuration files as described earlier.
  2. Ensure that PACKAGE_DIR, including the installer package, the certificates folder, config.yml, rpa.yml, ports.yml, and hosts.yml, are synchronized across all Linux hosts.

Back up old installation directory​

The procedure is optional.

To back up the old installation directory:

  1. Connect to each Linux server via SSH and, on each of them, run the following commands as APPLICATION_USER to create a backup directory:

    $ mkdir <PATH_TO_BACKUP_DIR>

  2. Connect to the server containing the new installation package via SSH and run the following commands as APPLICATION_USER to back up the old installation directory (INSTALL_DIR):

    $ cd INSTALL_DIR/wf_installer_new
    $ export ANSIBLE_PRIVATE_KEY_FILE="<path_to_ssh_key>"   # Path to the SSH key for connection to all Linux servers as APPLICATION_USER
    $ export ANSIBLE_REMOTE_USER="APPLICATION_USER"                # Application user specified in config.yml
    $ export ANSIBLE_VAULT_PASS="<ANSIBLE_VAULT_PASSWORD>"  # Password for decrypting config.yml
    $ ./install.sh backup full -e backup_dir=/PATH_TO_BACKUP_DIR # Path to the directory to place a backup. Make sure it is writable for APPLICATION_USER.
note

The preceding procedure is intended for the single-point installation. To back up the directory created during the multi-point installation, specify <SERVER_ROLE> instead of full in the previous command, for example:

$ ./install.sh backup <SERVER_ROLE> -e backup_dir=/PATH_TO_BACKUP_DIR -c local --limit <CURRENT_HOSTNAME> # here, <server_role> may be int, app, bepagent, etc.

Upgrade components on servers​

caution

Before uninstalling the RPA component, manually copy the log content to preserve the logs. The corresponding folder and all of its content will be deleted.

Log locations:

  • bot-agent logs: ..\RPA\bot-agent\logs
  • worker logs: ..\RPA\logs
  • nginx logs: ..\RPA\nginx\logs
  • WFSvc logs: ..\RPA\tools\logs

Upgrade the components, depending on your installation approach:

Remember to run the following command on the Master server as APPLICATION_USER.

  1. Go to INSTALL_DIR/wf_installer_new:

    $ cd INSTALL_DIR/wf_installer_new

  2. Set up APPLICATION_USER:

    $ export ANSIBLE_PRIVATE_KEY_FILE="<path_to_ssh_key>"   # Path to SSH key for connection to all Linux servers as APPLICATION_USER.
    $ export ANSIBLE_VAULT_PASS="<ansible_vault_password>"   # Password for decryption of config.yml.

  3. Remove outdated components of the previous installation:

    $ cd PACKAGE_DIR
    $ ./install.sh cleanup rudiments

  4. Run the following commands:

    $ ./install.sh precheck mssql   # Checks that MSSQL credentials are set correctly in config.yml, Database exists, wf_dba_user exists and has db_owner role
    $ ./install.sh configure mssql  # Creates all required schemas in Workfusion DB, creates and maps DB users to provided MSSQL logins, and assigns correct user permissions per schemas

  5. Run the upgrade commands:

    $ ./install.sh preinstall full -e skip_bi=true -e skip_rpa=true -e skip_ocrwin=true
    $ ./install.sh install full -e skip_bi=true -e skip_rpa=true -e skip_ocrwin=true
    $ ./install.sh check full -e skip_bi=true -e skip_rpa=true -e skip_ocrwin=true

  6. Optional.

    For DR installations: if you are updating a DR Failover environment, execute steps 5-8 from the Deployment section.

  7. Before updating the RPA server, uninstall the current RPA and perform a clean installation of a new RPA:

    $ ./install.sh uninstall rpa
    $ ./install.sh preinstall rpa
    $ ./install.sh install rpa
    $ ./install.sh check rpa

  8. Cleanup RabbitMQ queues:

    1. On the server with the installation packages, run the command:

      $ ./install.sh stop_services full # stop all services

    2. On the server with RabbitMQ (the first Master server in the HA mode or the Master server), run the commands:

      $ wfmanager start rabbitmq
      $ rabbitmq-diagnostics ping -q # repeat untill Ping succeeded (echo $? == 0)
      $ rabbitmqctl --silent list_queues --vhost bep | awk '{ print $1 }' | xargs -r -L1 rabbitmqctl delete_queue --vhost bep

    3. On the server with the installation package, run the command:

      $ ./install.sh start_services full # start all services

The upgrade is complete. It may take about five minutes for Control Tower to start. Wait several minutes if you see error 503 in a browser right after the upgrade.

Upgrade Analytics server​

Back up site configuration​

Before upgrading the Tableau Analytics components, connect to the Analytics (BI) server via RDP as a Windows Administrator user and save all your custom workbooks and data sources to a predefined location on a local disk.

info

Download custom workbooks and datasources to separate directories. For example, c:\patch_10_1_6_backup\workbooks\ and c:\patch_10_1_6_backup\datasources\.

Don't save backup files in the Analytics server installation directory, for example, c:\workfusion\, as it gets deleted during the upgrade.

To download the custom workbooks from the Analytics server interface:

  1. Go to a particular site. On the Content tab, click Workbooks. Look for the _custom postfix in their names to identify custom workbooks.

  2. In the right corner of the Name column, click More (...), click Download, and save the file as a Tableau workbook.

To download the custom data sources from the Analytics server UI:

  1. Go to a particular site. On the Content tab, click Datasources.

  2. In the right corner of the Name column, click More (...), click Download, and save the file as a Tableau datasource.

  3. Remember to rename the custom dashboards by replacing all occurrences of the underscore character ("_") with the plus symbol ("+") in the titles, except for the first and last two underscores.

  4. Follow the instructions in the Backup and restore custom Tableau dashboards section to restore custom dashboards.

Upgrade Analytics​

To upgrade the Analytics Server, follow the steps below:

  1. On the Master server, run the following command:

    $ cd /opt/workfusion/wf_installer_new
    $ ./install.sh uninstall bi

  2. Restart the Analytics server to apply changes.

  3. On the Master server, run the following command:

    $ ./install.sh preinstall bi
    $ ./install.sh install bi
    $ ./install.sh check bi

Restore custom parameters from previous versions​

note

When upgrading to v.10.2.8.x, you must manually configure LDAP/AD and IDP/SSO in Keycloak.

Custom workers​

Suppose you have changed workers' settings in the platform v10.1.x, you can migrate them after the upgrade.

After upgrading, on the Master server, open the file with workers' settings for Control Tower, for example, /opt/workfusion/vds-data/workers/app/com.workfusion.spa.ct/worker-app/10.1.0.32/worker.yml, and enter the previous corresponding settings, such as memory, CPU, and so on. Note that the structure of the new file differs from the previous one.

worker.yml for 10.1.x:

#cat /opt/workfusion/vds-data/workers/app/com.workfusion.spa.ct/worker-app/10.1.0.10/worker.yml

command: ${java} -XX:CICompilerCount=2 -XX:+UseSerialGC -XX:MaxMetaspaceSize=${metaspace}M -Xmx${heap}M -XX:+ExitOnOutOfMemoryError -Dgroovy.use.classvalue=true -Dgroovy.target.indy=true -Dfile.encoding=UTF8 -Dexecution.task.queue.input=${tasks.queue} -Dlogging.path=${log.dir} -Dlogging.file=${log.file} -Djava.io.tmpdir=${temp.dir} -Dwebharvest.machine.config.bundle.dir=${working.dir}/bcb -jar ${worker.jar} ${config.server.parameters}
cpu: 0.25
memory:
  heap: 1024

worker.yml for 10.2.8.x:

# cat /opt/workfusion/vds-data/workers/app/com.workfusion.spa.ct/worker-app/10.1.0.32/worker.yml

command: ${java} -XX:CICompilerCount=2 -XX:+UseSerialGC -XX:MaxMetaspaceSize=${metaspace}M -Xmx${heap}M -XX:+ExitOnOutOfMemoryError -Dgroovy.use.classvalue=true -Dgroovy.target.indy=true -Dfile.encoding=UTF8 -Dexecution.task.queue.input=${tasks.queue} -Dlogging.path=${log.dir} -Dlogging.file=${log.file} -Djava.io.tmpdir=${temp.dir} -Dwebharvest.machine.config.bundle.dir=${temp.dir}/bcb -jar ${worker.jar} ${config.server.parameters}
cpu: 0.4
memory:
  heap: 1024
  metaspace: 512
health-checks:
  - protocol: MESOS_HTTP
    path: /actuator/health
    grace-period-sec: 180
    interval-sec: 60
    timeout-sec: 10
    max-failures: 3

Upgrade Business Processes​

Migrate your Business Processes with Manual Tasks that use qualifications. To do that, refer to the following guides:

Check upgrade​

Once the upgrade is complete, check the WorkFusion Platform with Business Process (BP) and Manual Task. The OCR must be activated. For more information on post-installation checks, see Run post-installation check.

On the DR Failover environment, disable autostart of its processes. For that, execute step 11 from the Deployment section.