S3 (and compatible) backup target.

PradeepC

How will this feature work?

Can we decide only one backup target or multiple backup targets are allowed?
Particular server in Australia can be assigned with a backup location in the nearby DC in Australia, and for a server in Europe a different backup location can be added near Europe.

This will enable faster recovery of the backups when needed and also might save bandwidth costs

Aliysa_Enhance

The S3 compatible backup will allow you to select 1 backup target for now. As part of ongoing improvements this may be expanded to allow overrides per server.

You can run the built in Enhance backups and S3 backups side by side. The S3 backups will adhere to the backup role settings in Settings > Service > Backups.

Zoinkies

Aliysa_Enhance when will s3 backups be released?

hwcltjn

Aliysa_Enhance
Just to clarify, does that mean we can only add 1x S3 bucket as a backup location for the whole cluster?

PradeepC

Zoinkies Looking at the roadmap, I think it should be released by the end of this month.

Aliysa_Enhance

hwcltjn That's correct

hwcltjn

Aliysa_Enhance

0o, any particular reason why?

Adam

hwcltjn

This was simpler for the initial release. We may be able to offer more granular options but the issue is that the backups belong to the website and not the server that they are on; if the server changes you will want the backups to remain where they are.

Powie Do we still need a dedicated backup server with btfs storage or is it possible to use ONLY s3 ?

You can run only S3 however Enhance backups are a lot more space efficient and faster to create and restore. S3 backups are mainly intended for customers with smaller clusters who do not want to have a dedicated backup server or VM.

Powie

Do we still need a dedicated backup server with btfs storage or is it possible to use ONLY s3 ?

jayjayuk

Adam

What about restoring the sites to another server?

e.g. you run 1 server for everything and it fails, can we just spin up a new one. Show it the S3 bucket and restore and away we go?

Adam

jayjayuk

If you lose a server within your cluster, you can restore the sites from the S3 backup data to another server in the same cluster just as you can with the normal backup system.

It's not possible yet to restore the S3 backups to a new cluster/control server yet. When we release the "downloadable backups" feature this will be possible.

Aliysa_Enhance

S3 support is now live in 9.10.0 (https://enhance.com/support/release-notes). We've also created a guide on how to get started using Backblaze (https://enhance.com/docs/backup-role/s3-backups.html). I know this is one of many providers, so feel free to comment below if you'd like us to add a guide on your preferred provider.

Any problems/ questions please don't hesitate to get in touch.

jeff

So my understanding based on what I've seen posted is that it's either:

1: Use a backup server
OR
2: Use and S3 endpoint?

Is that correct?

Adam

You can use both simultaneously if you want to. The automatic backups will be made to both S3 and the Enhance backup server. The manual backups will prefer S3.

FranzL

So we still need a backup server correct? We just now have the option to store locally or remote or both?

Adam

You can choose to run either:

Enhance backups only (recommended for larger clusters)
S3 backups only, no backup server required (recommended for smaller/single-server clusters)
Enhance backups and S3 backups (for extra redundancy)
No backups (maybe if you only host testing/development sites)

FranzL

Adam

Adam S3 backups only, no backup server required (recommended for smaller/single-server clusters)

I added the S3 connection but If I try to edit a package the backups section is still locked for me. No backup servers exist in the cluster.

george

Adam Enhance backups are super stable and reliable!
Zero issues and I am extremely satisfied.

SMO

FranzL Same here.

Abdeen

Adam Is it possible for S3 backups to be incremental?

FranzL Same.

Adam

We think this is a UI bug. The backups will still be taken, it's just not possible to enable it on a package level if you don't have any backup servers installed on your cluster. We will shortly release a patch for this, as well as some other bug fixes and improvements to S3 backups.

FranzL

Adam

Adam The backups will still be taken, it's just not possible to enable it on a package level if you don't have any backup servers installed on your cluster.

Yes I saw that they ran overnight.

Aliysa_Enhance

Abdeen It might be possible in the future, at the moment S3 backups generate a full tar.gz of each website.