Kasten + Veeam Repo

[mkretzer]

Hello,

we think about integrating Kasten K10 with our Veeam Backup Infrastructure.

According to https://www.veeam.com/kb4258 "Restore points can be offloaded to tape for longer retention. However, after restore, only Veeam-based recovery options can be used with these points".

Do i understand it correctly that if for example we need to access deleted data of an elasticsearch cluster, backed up by Kasten from 4 weeks after deletion we cannot simple restore an old tape backup from tape to repo and then use Kasten to restore the elasticsearch cluster?

Markus

[bct44]

Hello,

For me your logic is good and the use case is real. Unfortunately i'm waiting some network works to test it on my infrastructure.

If you want an answer before testing, i will try to open a support case to kasten team or post your question on the k10 support group on the community hub.
https://community.veeam.com/groups/kast ... support-92

May i ask you? Why do you want to use kasten for elasticsearch workload? You can just send a snapshot to an s3 endpoint and restore it if needed. Is it part of your backup strategy for an entire app?

[mkretzer]

We have ~80 k8s clusters with hundreds of nodes and each has on average about one Elastic Instance, some multiple. Each one host a very different environment and the application is created by different dev teams. The IT team managing the k8s/backup/infrastructure consists of about 3 engineers, of which only one does this as his primary job. So we need a solution which takes as much work off our shoulders, so we do as much automation as possible. Right now we have not yet started using Kasten, but back up the NFS elastic is writing to, which means restores are really hard (but work).
My ideal solution would be something we can deploy to all clusters, which auto-detects all services that can be backed up and allows to manage this in a central console for backup.

[bct44]

That makes perfect sense, i hope some guys from kasten/Veeam PM could give an answer regarding your interesting use case.

[Lei.Wei]

Hello Markus
In this integration, K10 only sends PVC data to the VBR repo, so you need to store all the YAML metadata in your S3 bucket. When you do K10-based recovery, you need restore points from both repositories. So if your k10 retention policy removes your restore points after VBR offloads these points into your tape, you may not see these restore points from K10 GUI. That's why only Veeam-based recovery options can be used from the VBR console.

[mkretzer]

Hello,

could we use the new S3 Backup capability of 12.1 to backup the Data from S3 to Repo to tape and back?

Markus