Commented by: @dyemanov

What architecture is affected? Do you have a test case? I tried restoring various backups with SS and SC and got no problems.

Commented by: @AlexPeshkoff

On linux (both SS/CS) I've also failed to reproduce a bug.

Commented by: @alexpotapchenko

I built server v. 2.5.0.19089 from cvs. This error is reproduced in all architectures with my private database backup (~70 mb). If you can't reproduce this error on a backup of the similar size, I will try to make the test case.

Commented by: @dyemanov

The biggest backup I tried was 280MB but still without luck.

Commented by: @alexpotapchenko

I reproduced this bug with backup in attachment.
I used command: gbak.exe -c c:\test.fbk c:\test.fdb -user sysdba -password masterkey -se localhost:service_mgr -v

Commented by: @AlexPeshkoff

I've reproduced some problems, related with attached database backup.

Commented by: @dyemanov

Me too. One issue is related to the not found partner index (the FK code still seems buggy), another one is an assert with an incorrect status vector.

Commented by: @AlexPeshkoff

Alexander, can you comment please - how did it happen that absolutely all indices (including primary keys) are marked for deferred activation in your backup? If this is due to manual modification of system tables, it's OK. But if not, this means one more bug, and I need to know - how was this backup created?

Commented by: @AlexPeshkoff

Direct reason of a bug was passing an error in status vector from service even if it processed that error itself. But it also helped to solve a lot of problems with actively running MT engine shutdown.