I’m looking to update my ghost instance. The documentation states that I can backup my data with ghost backup. However, it never states how to use this backup data to restore my previous ghost instance quickly in case things go wrong. How would I be able to do it?
See the following. The first explains that Ghost backup takes a copy of your data (toward the end of the page.) The second, how to restore from a manual backup.
You also have the option to use ghost update --rollback if an update is unsuccessful.
Hi mjw, thank you for the explanation. I understand now that you’ve got an automated way of rollbacking an update, which is reassuring. I also understand that you probably recommend keeping a backup of my data because these rollbacks might fail - which is normal.
I think the documentation could be slightly improved, if I may suggest:
“a copy of your data” is not a precise statement. It’s probably not a copy of the ghost entire directory, otherwise you would not need to build a specific command for it. I’ve also recently backed up and restored (manually) an instance of ghost, which worked for the most part but broke most of the images. I think that rather than describing every piece of data, it may be more useful to describe somewhere how to use this backup to restore a ghost instance fully. If I see a ghost backup command, I expect that there is a corresponding ghost restore command. Since there is no ghost restore, I’m left wondering what piece of data might be missing from this backup. In short, my suggestion is to add somewhere what we should do with this zip in case things go awry - it’s quite unclear for now.
The rollback mechanism should be mentioned on the Ghost Docs page. On an update documentation, I always wonder “what if things go wrong, what do I do?” and going to the CLI full list of commands is not my initial reaction.
I hope this helps improve the product, which I really love. I’ll get started on my upgrade to latest now :)
I disagree. Ghost backup takes a copy of the contents folder, an export of the database plus routes and redirects. You can reconstruct your site with these data.
This isn’t the purpose of the command. If you want to recover in this way, use a system-wide backup solution or snapshots.
Your explanation about “a copy of your data” here is more helpful and precise. You’re actively shooting down my ideas though, I’ll refrain from giving more feedback in the future. I now love the product a little less