MarkLogic - API Request & Restore Database

Question

We are using MarkLogic 9 version. We have developed API above MarkLogic

We are doing daily full backup for last 7 days.

Now we are upgrading our MarkLogic server to 10 & as a part of disaster recovery if due to some reason upgrade fails and we need to restore backup from yesterday

I want to understand

1. How restore process works, do restore for each stand and then serve from remaining ones?
2. Whether API requests will be served during restore process ?
3. If API requests will be served then which data will be used to serve that request ?
4. Do we need to go for downtime as a part of restore process ?
5. If we go for incremental backup & then restore, will there be any difference to above points?

Answer 1

The Restore will restore the entire backup to a temporary directory. Once all of the stands/forests have been copied to disk it will swap the current forests/stands with the restored ones.

API requests will still be serviced by the active stands, with the existing data. It will not use the restored data until after the restore completes.

There will be a very small amount of downtime as the existing forests shutdown, and the restored forests start up. This is usually just a few seconds, but it does depend on how big the forests are.

There is no difference in the behavior if you are restoring a full + incremental/s or only a full backup.

Be aware that you will need enough disk space for both the current data and the restored data, as they will coexist for a period of time.

Backup and Restore Transactions

Phases of Backup or Restore Operation