Migrating Kyvos from CentOS to Ubuntu on GCP
This documentation is for Kyvos 2026.8.x. To view documentation for other Kyvos versions, click here. |
|---|
1 Overview
This section describes how to migrate Kyvos instances from CentOS to Ubuntu on Google Cloud Platform (GCP). The migration is performed by replacing the boot disk on each machine and updating the associated startup scripts and configurations.
2. Prerequisites
Before you begin the migration, ensure the following:
Kyvos and Kyvos Manager are not installed on the root drive.
Capture the Kyvos Manager path while it is still running, since you will need it later to restart Kyvos Manager. Run:
ps -ef | grep 8080
On this page:
3. Migration Steps
Step 1: Prepare the Ubuntu Boot Disk
From the GCP Console, click Activate Cloud Shell.
In Cloud Shell, run the following commands, replacing
{REGION}and{ZONE}with the appropriate values for your environment:gcloud config set compute/region {REGION} gcloud config set compute/zone {ZONE} gcloud compute disks create ubuntu-boot-disk \ --image projects/dev-enterprises-273509/global/images/kyvos-ubuntu20-image-2023-11-07 gcloud compute snapshots create ubuntu-boot-disk-snap \ --source-disk-zone={ZONE} \ --source-disk=ubuntu-boot-disk \ --snapshot-type=STANDARD
Note
This creates the Ubuntu boot disk and a snapshot of it, which will be used to configure the boot disks on your Kyvos machines in later steps.
Step 2: Migrate the Analytical Server and Query Engine
Note: Repeat steps 1–7 below for both the Analytical Server and the Query Engine machines.
From the Google Cloud Console, stop the Analytical Server / Query Engine machine.
Once the machine is stopped, click on it to open its details screen, then click Edit.
Scroll down to the boot disk section and click Detach Boot Disk. Once detached, the option to configure a new boot disk becomes available.
Click Configure Boot Disk. In the side panel that opens:
Select Snapshots.
Choose the snapshot created in Step 1 (
ubuntu-boot-disk-snap).Provide a name of your choice.
Set the device name to
boot.Leave the remaining fields as auto-filled.
Scroll to the Startup script section:
Replace the existing content with the new startup script: bi_qe.sh
In the metadata, add a key named
Rerunwith the valuetrue. If theRerunkey already exists, update its value totrue.
Click Save.
Verify that the startup script content was saved correctly. Do not proceed until this is confirmed.
Start the Analytical Server and Query Engine machines.
On the Kyvos Manager Dashboard, verify that the Analytical Server and Query Engine are up and running.
Step 3: Migrate the Kyvos Manager
From the Google Cloud Console, stop the Kyvos Manager machine.
Detach the old boot disk and attach a new disk of the same size, following the same process used for the Analytical Server and Query Engine machines (Step 2, sub-steps 2–4).
Update the
Rerunmetadata value fromfalsetotrue.Remove the existing startup script content and replace it with the new script: km.sh
Click Save.
Verify that the startup script content was saved correctly. Do not proceed until this is confirmed.
Start the Kyvos Manager machine.
Restart Kyvos Manager using the path captured in the Prerequisites step, and verify that all machines and services are up and running.
Step 4: Update PostgreSQL Connection Script
Replace the
pgconnect.shscript located inpostgres/binwith the updated version: pgconnect.sh
4. Verifying the OS Migration
To confirm that an instance has been successfully migrated from CentOS to Ubuntu 20, run the following command on the Kyvos component:
lsb_release -aSample output:
Distributor ID: Ubuntu
Description: Ubuntu 20.04.6 LTS
Release: 20.04
Codename: focal5. Post-Migration Cleanup
Once the OS upgrade has been completed and verified on all instances, delete the boot disk and snapshot created in Step 1 (
ubuntu-boot-diskandubuntu-boot-disk-snap).