Snapshot issue

Elastic Search
1

hi all

I wrote a policy to take a snapshot daily and put the data in the defined path.repo. when I check the list of the taken snapshots using below api, I can see the list of snapshots based on the defined policy's naming method, therefore it is easy to find the desire snapshot .

GET /_snapshot/my_fs_backup/_all

but when it comes to take a backup from the path of repository to another storage ( tape for example) , we will face the problem, because naming in the repository folder is quite different from what we see in snapshot list in dev tools. Indeed, it is desired to take a backup from repository folder using a scheduled job based on the date whereas there is no option to set dynamically path.repo based on the date. Is there any way to resolve this issue?
Any help well be so appreciated

2

If you take a backup of the path of the repository and want to restore it, you would need to add a new repository in Elasticsearch pointing to this path.

Also, while making a backup of a repository you need to follow these instructions to avoid corrupting it.

3

I'm reading @sahere37 Q a bit differently. Perhaps he/she can clarify.

It's also not completely clear what's being snapshot-ed to the repo. Each day a different index is snapshot-ed? or a data stream? Or the full cluster?

My reading is that they don't really know which files in the repo (filesystem) are from which snapshot. Which is right of course, it's far from obvious. So they are unable to make a (eg tape) backup of "this specific snapshot" or indeed restore "that specific snapshot" from tape. Maybe for perceived speed of recovery (from tape) reasons.

Maybe I read the Q wrong of course.

4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.