Offline upgrades

To upgrade Plixer Scrutinizer Collectors/servers that do not have access to https://files.plixer.com, an offline repository will need to be created. This involves downloading the offline repository file for Plixer Scrutinizer and moving it to the primary Reporter.

Important

When upgrading from v18.20 to v19.3.2, a separate Plixer Scrutinizer v19.3.2 virtual appliance will need to be used as the offline repository. The primary Reporter in a v18.20 cluster cannot be used for this purpose. Once the upgrade has been successfully installed, the virtual appliance can be safely deleted. For assistance, contact Plixer Technical Support.

To set up the offline repository and start the upgrade process, follow these steps:

1. Perform a backup of the current Plixer Scrutinizer install as described in the recommended upgrade preparation steps.

2. Download https://files.plixer.com/plixer-repo/scrutinizer/19.3.2_offline.tgz to a computer with Internet access.

3. Download https://files.plixer.com/plixer-repo/scrutinizer/19.3.2-checksum-offline.txt and validate the checksum of the offline repository .tar file.

4. Enter df -h and confirm that the host to be used for the offline repository has at least 84 GB of free disk space under /var/db/big (hardware appliances) or /var/db (virtual appliances).

5. Start an SSH session with the offline repository host as the plixer user.

6. Create a new directory for the offline installation files and set the correct permissions to give the plixer user access to it.

   For hardware appliances:

   sudo mkdir /var/db/big/offline
   
   sudo chown plixer:plixer /var/db/big/offline
   

   For virtual appliances:

   sudo mkdir /var/db/offline
   
   sudo chown plixer:plixer /var/db/offline
   

7. Move the .tar file to the appropriate directory on the offline repository host.

   For hardware appliances:

   scp 19.3.2_offline.tgz [email protected]:/var/db/big/offline/19.3.2_offline.tgz
   

   For virtual appliances:

   scp 19.3.2_offline.tgz [email protected]:/var/db/offline/19.3.2_offline.tgz
   

8. Extract the contents of the file to the same directory.

   For hardware appliances:

   sudo tar -zxvf /var/db/big/offline/19.3.2_offline.tgz -C /var/db/big​/offline
   

   For virtual appliances:

   sudo tar -zxvf /var/db/big/offline/19.3.2_offline.tgz -C /var/db/offline
   

   Important

   On 19.2.0 virtual appliances, extracting the contents of the offline repository tar file will exhaust all available space on the template machine.

9. Create a symlink to the offline directory from the html directory, so the files can be served by Apache.

   For hardware appliances:

   sudo ln -s /var/db/big/offline/plixer-repo /home/plixer/scrutinizer/html/plixer-repo​
   

   For virtual appliances:

   sudo ln -s /var/db/offline/plixer-repo /home/plixer/scrutinizer/html/plixer-repo​
   

Important

In the two following steps, replace REPO_HOST with the IP address or hostname of the offline repository host.

10. From the primary Reporter, download the v19.3.2 installer from the offline repository host.

curl -o scrutinizer-install.run -L -k​ https://REPO_HOST/plixer-repo/scrutinizer/19.3.2/scrutinizer-install.run​

11. Download the checksum file and validate the integrity of the scrutinizer-install.run file.

curl -o scrutinizer-checksums.txt -k https://REPO_HOST/plixer-repo/scrutinizer/19.3.2/scrutinizer-checksums.txt

cat scrutinizer-checksums.txt

sha256sum scrutinizer-install.run

12. Set the correct permissions for the installer.

sudo chmod 755 scrutinizer-install.run

Important

For AMI cluster upgrades, it is also necessary to run chmod 755 /home/ec2-user before starting the installer. This will ensure that the SSH key used for communication between Plixer Scrutinizer Collectors has the correct permissions. Contact Plixer Technical Support for assistance.

13. Run the scrutinizer-install.run file as the root, plixer (recommended), or ec2-user user. x.x.x.x should be replaced with the IP address or hostname of the offline repository host. The -- -k argument is used to ignore certificate errors.

REPO_HOST=x.x.x.x ./scrutinizer-install.run​ -- -k

Important

When asked whether to reboot automatically after the upgrade, enter n. This will prevent distributed cluster nodes from rebooting before the primary Reporter has completed upgrading.

Note

If the server was previously upgraded from 18.20, you will be asked whether to delete the data.old backup that was created during that upgrade. Because a new backup should have been created prior to starting the current upgrade process, data.old can safely be deleted.

14. Distributed cluster upgrades only - When asked how the installer should log in to remote Collectors in the cluster, enter either existing (recommended) or passwords.

Important

When upgrading a distributed cluster from v18.20, passwords should be used. Prior to this step, the installer will also prompt the user to create a new plixer control key, which should be left blank unless ecnrypted keys are required.

15. After the installer has finished running, the following heartbeat checks should be run to verify that all nodes (including standalone deployments) are able to communicate normally:

scrut_util --check heartbeat --type database
scrut_util --check heartbeat --type api

If all heartbeat checks are successful, then the systems have been upgraded successfully.