Expand my Community achievements bar.

Learn about Edge Delivery Services in upcoming GEM session
SOLVED

Incremental content backup

Avatar

Level 2

Incremental content backup

I have been going through the docs, posts and other external sources to find a way to achieve the below: 

a) I want everything under /content/site and few other nodes say: /etc/designs/site, to be incrementally backed up in frequent intervals - creating incremental hourly backups of content and packaged up (zip files).

b) the hourly backups created in the past 24-hours would be recycled everyday, once the first hourly backup is cut for the day. 

I remember having had to deal with this scenario in the past (may be an year or two ago), and there was no straightforward way to do it other than writing custom application to create packages in this fashion with incremental changes. If there is nothing available out of the box with package manager (looks like there isn't), this could be good candidate for feature request.

We are expecting a lot of content updates on a daily basis, and cannot afford to lose any content - reason why we would need something more than nightly full repo, system wide backups. 

I would like to know how others are dealing with these or similar requirements. 

Best,
Sivaram

1 Accepted Solution

Avatar

Correct answer by
Level 8

The most common solution I have implemented to address the need for frequent back ups is to leverage an active - passive author cluster with a SAN data storage layer. This enables several scenarios, the common one however is:

This approach means that recovery is more complex because you have to recover and entire instance, you can't just install a package, however it's most practical way I know of to get frequent backups. 

View solution in original post

2 Replies

Avatar

Employee Advisor

In your case I would approach this differently. Because such an incremental backup is quite expensive (check all nodes/pages if they have changed and add them to a package definition), you should try to come up with a faster and less disruptive backup solution. If you have a good storage system you could use filesystem snapshots and run a backup every hour.

Jörg

Avatar

Correct answer by
Level 8

The most common solution I have implemented to address the need for frequent back ups is to leverage an active - passive author cluster with a SAN data storage layer. This enables several scenarios, the common one however is:

This approach means that recovery is more complex because you have to recover and entire instance, you can't just install a package, however it's most practical way I know of to get frequent backups.