Hi!

When im upgrading a physical UVM appliances i create a new backup from ip-or-domain/appliance

Also you can download an exe with your current version if you need to do rollback. To do this find the offline tool on C:/programfilex86/updater/service/

If you need more info about offline tool let me know it!

Hope this help you.

Hey, to add upon what l3g0l4s stated, our KB articles for upgrades in Active/Active or Active/Passive describe how to get the backup of the appliance page. That said, taking a backup snapshot is highly recommended as well before the upgrade. 

KB articles: 

• KB0017028 - Upgrade BeyondInsight Password Safe in an Active Passive Appliance setup

  • https://beyondtrustcorp.service-now.com/csm?id=kb_article_view&sysparm_article=KB0017028

• KB0017029 - Upgrade BeyondInsight Password Safe in an Active Active appliance setup

  • https://beyondtrustcorp.service-now.com/csm?id=kb_article_view&sysparm_article=KB0017029

*Note about viewing KB articles:

1. The links will work if you’re already logged into the customer portal, otherwise you’ll receive a permission denied
2. Alternatively, you can search for the articles in Community general search (as it will search public + internal) and bring you over to the KB article.

yes I am aware of how to take the backup via /appliance but the once BI or appliance management is upgraded, then the backup shows as legacy right? and restore doesn't work. So how can we utilize the backup taken? the KB articles doesn't mentioned anything on the recommended possible roll back processes for physical appliances . Had gone through those already

Thank you for your response. So we can do rollback to previous version by running the exe for previous version? wasn't aware of that..so DB schema changes etc will reflect the previous version once that installer is run, then we can restore via the backup as well?

Hey @SFA - fair point! If you’re at the point of doing a restore from an upgrade using the backup generated, you’ll likely be working with support (please!) and they’ll be guiding through situation specific restoration steps depending on what broke/triggered of the roll-back. 

Edit: Checking in with support pre-upgrade is also an option in case there are environment-specific questions that exist that can help ensure a successful & safe upgrade.

Sorry, but i cannot confirm it,  im not sure if will works, but i would do it.

To be sure maybe you can do a test on a virtualmachine environment.
1.- Buckup /appliance
2.- Upgrade
3.- Download previus version .exe and execute on updates folder.
4.- Restore the backup if it is necessary

In addition maybe some BT Engineer can help with this idea. Im interested in this topic because i work with physical machines too.

There isn't a documented roll back process for physicals via older exe.

For virtual, we take a powered off vm snapshot before any activity and roll back via the vm snapshot.

For physicals, the recommendation is to contact support to first review the situation, and to confirm the issue cannot be resolved, before starting the roll back. Contact support when observing any issues, do not attempt to try any manual fixes. Search KB for known issues.
 

Roll Back for physical:
- Verify again the last good backup is healthy.
- Rebuild and deploy a new image using usb and update it to the same version. generic rebuild guide.
- Restore the last good Backup.
- Verify data and application.

In some environments, this can a very lengthy process and we can reduce the timing when we have a ready Cold Spare that we can simply bring online quicker by just doing a restore and deactivate the cold spare role. We can also attempt to reduce or avoid issues in production by testing the activity in lower environments.

I always prepare with

stopping the password change agent

UVM based backups for integrated SQL or DB Cluster backups for Active/Active

Cold/offline snapshots in the hypervisor

Roll-back/back-out I always list as “fix on fail” which is primarily logging a severity 1 support ticket.

I have successfully created a test environment by cloning a UVM into a private network segment along with some cloned managed assets for testing; but this is a lot of work.