How to downgrade FMS and deploy a backup to the current DB

hello community, hello Todd,

On July 17, 2025, we compressed our database in the production environment, then migrated the development environment database to the production environment, and then upgraded Filemaker Server to version 22.0.1.66. Since then, we have been experiencing crashes of the fmserverd process. But only in the production environment, presumably only under load.

My plan now is to downgrade to FMS 21.1.4.400 and migrate the backup from before the compression to the current DB.

The goal is to fill the structure of the old database before the compression to FMS 21.1.4.400 with the data from the current DB.

Will that work?

We recovered the current DB. Only one invalid Relationship to a deleted TO was found. The recovery.log stated that the invalid relationship had been deleted, but it was still present in the restored database. However, I assume that the database is not the problem.

I will start in 3 hours.

Thanks in advance!

Hello,

I think you are asking if you can import data from the current file into a backup. If so then the answer is yes, that will work fine.

But if you did a recover and it found corruption then that corruption is likely in the backup also. But it may be the best you can do.

Hope that helps

Thanks

Todd

Hello,

just wanted to let you know, that we downgraded our servers and started the databases as they were (without migrating) and it works fine. No more crashes. Seems, the current FM server version has bugs. We haven’t received any response to the case we opened at claris support yet.

After that all I tried a deployment today but it got stuck at or after “build finished pushing”. I find the build in the outbox of the source server and an empty inbox at the target server. No errors in otto logs on both. DB on target server was closed.

I couldn’t abort the deployment. I had to stop the ottofms service on linux CLI.

Does anyone has an idea what is going wrong?

Hello,

The debug info log should show you something about what happened. It combines all the logs so they line up in order. Have you tried that.

Todd

unfortunately there’s no error. The last entry at the deployment log is
deployment_log: 2025-08-06T02:22:17.201Z info fetching Build finished pushing
and 2h later as I killed the otto service:
deployment_log: 2025-08-06T04:35:55.765Z error unknown phase Deployment crashed

I migrated another DB successfully. So basically ottofms works.

Hello, the deployment today worked fine, maybe ottofms wasn’t running properly before I restarted it yesterday.

Thank you very much anyway!

1 Like