# Not able to create schema while restoration in version 6.2.0

**URL:** <https://forum.scylladb.com/t/not-able-to-create-schema-while-restoration-in-version-6-2-0/4611>\
**Category:** ScyllaDB\
**Tags:** scylla-manager, backup-restore, upgrade, schema\
**Created:** [March 16, 2025, 8:50am UTC](https://forum.scylladb.com/t/not-able-to-create-schema-while-restoration-in-version-6-2-0/4611 "2025-03-16T08:50:41Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Shubham\_Jain](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.scylladb.com/shubham_jain/32/1584_2.png) [@Shubham\_Jain](https://forum.scylladb.com/u/Shubham_Jain)\
**Post date:** [March 16, 2025, 8:50am UTC](https://forum.scylladb.com/t/not-able-to-create-schema-while-restoration-in-version-6-2-0/4611/1 "2025-03-16T08:50:41Z")

</div>

_Installation details_  
Running a 3 node scylla cluster in GKE.  
#ScyllaDB version: 6.2.0  
#Cluster size: 3 node  
#os (RHEL/CentOS/Ubuntu/AWS AMI): Ubuntu

Hi all,  
I have upgraded my Scylla cluster from 5.2.4 to 6.2.0. now when I try to restore a schema from already running scylla cluster to newly created scylla cluster, using command:  
sctool restore -c scylla-backup -L gcs:scylla-backup-prod -T sm\_20250214180010UTC --restore-schema  
I’m getting issue that scylla-manager keyspace already exists. but when I delete scylla-manager namespace from new cluster, I can’t even run the above command.  
Earlier scylla used to store schema in .cql format but now it’s storing in json format.  
Any help on this is appreciated.  
Thanks!

---

<div class="post-metadata">

**Author:** ![Michal\_Leszczynski](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.scylladb.com/michal_leszczynski/32/429_2.png) [@Michal\_Leszczynski](https://forum.scylladb.com/u/Michal_Leszczynski)\
**Post date:** [March 17, 2025, 8:27am UTC](https://forum.scylladb.com/t/not-able-to-create-schema-while-restoration-in-version-6-2-0/4611/2 "2025-03-17T08:27:42Z")

</div>

Hi, could you share the scylla-manager logs and version?

Are you perhaps using the same cluster for your data and as the scylla-manager back-end? Have you created a new scylla-manager instance for this new cluster (and also using this new cluster as its backend)?

If that’s the case, then backup contains the schema definition of scylla-manager keyspace and the new cluster also contains such definition (created by the new scylla-manager on its start-up, as it is its back-end).

A simple fix is to not create a new scylla-manager, but use the old one for schema restoration in the new, fresh cluster. After schema and data restoration, for whatever reason, you can create a new scylla-manager instance with the newly restored cluster as its back-end, and all should work fine (although using the same cluster for your data and scylla-manager back-end is problematic as in this case).

Another option would be to restore schema by taking the .cql file from the backup and applying it manually (see [backup spec](https://manager.docs.scylladb.com/stable/backup/specification.html) for more info).