Migrating a KVM NFS-based Deployment via a Staging System

Migrating a KVM NFS-based Deployment via a Staging System

Overview

In this migration method, the production 3.7.x FortiSIEM systems are left untouched. A separate mirror image 3.7.x system is first created, and then upgraded to 4.2.1. The NFS storage is mounted to the upgraded 4.2.1 system, and the collectors are redirected to the upgraded 4.2.1 system. The upgraded 4.2.1 system now becomes the production system, while the old 3.7.6 system can be decommissioned. The collectors can then be upgraded one by one. The advantages of this method is minimal downtime in which incidents aren’t triggered, and no upgrade risk. If for some reason the upgrade fails, it can be aborted without any risk to your production CMDB data. The disadvantages of this method are the requirement for hardware to set up the mirror 3.7.x mirror system, and longer time to complete the upgrade because of the time needed to set up the mirror system.

The steps in this process are:

Overview

Prerequisites

Restoring the Upgraded CMDB in a 4.2.1 Virtual Appliance

Mounting the NFS Storage on Supervisors and Workers

Assigning the 3.7.x Supervisor’s IP Address to the 4.2.1 Supervisor Registering Workers to the Supervisor

Prerequisites

Contact AccelOps Support to reset your license

Take a snapshot of your 3.7.x installation for recovery purposes if needed

Make sure the 3.7.x virtual appliance has Internet access

Download the 4.2.1 migration scripts (ao-db-migration-4.2.1.tar). You will need the Username and Password associated with your AccelOps license to access the scripts.

Create the 3.7.x CMDB Archive

1. Log in to your running 3.7.x production AccelOp virtual appliance as root.
2. Change the directory to /root.
3. Copy the migration script ao-db-migration-4.2.1.tar to the /root
4. Untar the migration script.
5. Make sure that the owner of ao-db-migration.sh and ao-db-migration-archiver.sh files is root.
6. Run the archive script, specifying the directory where you want the archive file to be created.
7. Check that the archived files were successfully created in the destination directory.

You should see two files, cmdb-migration-*.tar, which will be used to migrate the 3.7.x CMDB, and opt-migration-*.tar, which contains files stored outside of CMDM that will be needed to restore the upgraded CMDB to your new 4.2.1 virtual appliance.

8. Copy the cmdb-migration-*.tar file to the 3.7.x staging Supervisor, using the same directory name you used in Step 6.
9. Copy the opt-migration-*.tar file to the /root directory of the 4.2.1 Supervisor.

Restoring the Upgraded CMDB in a 4.2.1 Virtual Appliance

1. Log in to your 4.2.1 virtual appliance as root.
2. Change the directory to /opt/phoenix/deployment/.
3. Run the post-ao-db-migration.sh script with the 3.7.x migration files phoenixdb_migration_xyz and opt-migration-*.ta r.
4. When the migration script completes the virtual appliance will reboot.

Mounting the NFS Storage on Supervisors and Workers

Follow this process for each Supervisor and Worker in your deployment.

1. Log in to your virtual appliance as root over SSH.
2. Run the mount command to check the mount location.
3. Change to the 3.7.x mount path location in the /etc/fstab file on the Supervisor or Workers.
4. Reboot the Supervisor or Worker.

Registering Workers to the Supervisor

1. Log in to the Supervisor as admin.
2. Go to Admin > License Management.
3. Under VA Information, click Add, and add the Worker.
4. Under Admin > Collector Health and Cloud Health, check that the health of the virtual appliances is normal.

Setting the 4.2.1 SVN Password to the 3.7.x Password

1. Log in to the 4.2.1 Supervisor as root over SSH.
2. Change the directory to /opt/phoenix/deployment/jumpbox.
3. Run the SVN password reset script ./phsetsvnpwd.sh
4. Enter the following full admin credential to reset SVN password

Organization: Super

User: admin

Password:****

Migration is now complete – Make sure all devices, user created rules, reports, dashboards are migrated successfully

Migrating Collectors

1. After migrating all your Supervisors and Workers to 4.2.1, install the 4.2.1 Collectors.
2. SSH to the 3.7.x Collector as root.
3. Change the directory to /opt/phoenix/cache/parser/events.
4. Copy the files from this directory to the same directory on the 4.2.1 system.
5. Change the directory to /opt/phoenix/cache/parser/upload/svn.
6. Copy the files from this directory to the same directory on the 4.2.1 system.
7. Power off the 3.7.x Collector.
8. SSH to the 4.2.1 Collector and change its IP address to the same as the 3.7.x Collector by running the vami_config_net
9. In a browser, navigate to https://<4.2.1_Collector_IP_address>:5480 and fill in the administration information to complete the Collector setup/

Migrating the SVN Repository to a Separate Partition on a Local Disk

If you are using NFS storage, your SVN repository will be migrated to a local disk to improve performance and reliability. If you are using local storage only, the SVN repository will be moved out of the /data partition and into an /svn partition.

3. Download ao-svn-migration.sh script from image server. (https://images.FortiSIEM.net/upgrade/va/4.3.1)
4. Copy or move the ao-svn-migration.sh script to /root.
5. Run ls -al to check that root is the owner of ao-svn-migration.sh.
6. Run chmod to change the permissions on ao-svn-migration.sh to 755.
7. Reboot the machine.
8. Log into the Supervisor as root.
9. When the script executes, you will be asked to confirm that you have 60GB of local storage available for the migration. When the script completes, you will see the message Upgrade Completed. SVN disk migration done.
10. Run df –h to confirm that the /svn partition was completed.

Special pre-upgrade instruction for 4.3.3

1. SSH as root into the Supervisor node
2. Download “phupdateinstall-4.3.3.sh” script
3. Copy or move the phupdateinstall-4.3.3.sh script to /root
4. Run chmod to change the permissions on phupdateinstall-4.3.3.sh to 755

Special pre-upgrade instruction for 4.6.1

Instructions for Supervisor node

Run the following command as root.

Instructions for Collector  nodes

Run the following command as root on each collector prior to upgrading the collector from the GUI, or the upgrade will fail:

Enabling TLS 1.2 Patch On Old Collectors

Older AccelOps collectors 4.5.2 or earlier running JDK 1.7 do not have TLS 1.2 enabled. To enable them to communicate to FortiSIEM 4.6.3, follow these steps

1. SSH to Collector and edit /opt/phoenix/bin/runJavaAgent.sh