Hi!! I want to know, it’s possible have to NCP instances running to get 100% up time if one fail? Or what you recomend to have a backup of a NCP instance to recover it quickly
Possible, yes, but that knowlegebase is subscriber exclusive content.
Having adequate backups of the instance is my strategy. Mine are set to auto using ncp-config but can achieve same through ncp-web.
That knowledgebase assumes you have many machines, not just two of them. It is for enterprise users running lots of servers.
Join us on https://t.me/nextcloudpi
I’d also like to set something up. Best I can see is…
- Device with NCP
- Device with Data, probably in NFS or Samba.
- You can use BTRFS (ncp method) or use an rsync tool like Restic (Rclone) to manage full backups of your NCP device.
- External Backup of Data
I know people use Docker Swarm, Kubernetes, GlusterFS and other tools… I’m not sure of how it could work with only 2 devices. Swarm, etc. will need at least 3 iirc.
I guess having a full failover system running which seamlessly takes over will (for most cases) be quite an overkill. It requires not only to keep the data in sync but also the SQL database - which is a lot more hassle. For me, the two options would be:
- As a private user with a cloud for family and friends, make sure to back up things regularly. Then, keep a second server around to which you can quickly deploy your backups (yes, backup up also means to have a way of restoring them… ) and manually switch your domain over to that system
- As enterprise user with mission-critical deployment, get a service subscription with the NC people and let them set up the thing for you properly.
I suggest this because I believe that you can do more harm to the stability, availability and data integrity of your server than actually win from the setup if not done properly.