Timescale DB single helm - how to do a point in time recovery with 3 node kubernetes cluster

i’m deploying timescale db single on a 3 node kubernetes cluster(self hosted) using your helm chart:

i was able to configure backup on AWS S3 bucket according to the Admin Guide:

i was abale to recover from master failure using the following steps:

  1. take down the cluster bu uninstalling the chart (helm uninstall)
  2. clear the data and wal directories(on local volumes)
  3. change values.yaml:
    3.1 add boostrap from backup configuration
    enabled: true
    repo1-path: /backup
    3.2 disable backup
    enabled: false
  4. redepoly the cluster again

everything works as expected!!

the question is how can i perform a Point In Time Restore from S3 Backup
i didn’t find any configuration i can use in the values.yaml
this is an option in case we have data corruption that started at some point in time, and for sure this data correption will be recovered in the normal bootstrapFromBackup, as i assume tha latest backup is used

what is the procedure to perform point in time recovery, as part of the helm deployment
or to perform restore from a specific full backup located on the S3 bucket


You can find the script that do the pgbackrest restore here.
Probably you can go to restore documentation on pgbackrest and change the code to restore from aws if necessary.

ok i will have a look thanks.

Is there any chance your deploy will include those options(Point In Time recovery) to be configured via values.yaml ?

I think you can do it via Patroni. Check more here.