Diego,

> Every week we make test of the restore procedure of our production
> databases.
>
> 3 weeks ago we tried to restore one of them as usual, using Ibexpert with
> verbose option to screen.
> The restore procedure writes all the data to the DB, but when it tries to
> create the indexes for the tables the process hangs without any message.
> Always in the same index.
> If we choose "deactivate indexes" in the restore procedure we can finish the
> restore, and after that we activate all the indexes with a script and we can
> finish the "restore" process.
>
> If we do a backup/restore of that restored db the process finish fine.
>
> We made those tests in a test server. I'm thinking to do the work around
> procedure of the restore to the production db to "fix" my issue.
> But I like to understand the reason of that problem.
>
> We are using FB 1.5 (We are planning to migrate to 2.1 or 2.5 as soon we have
> the new Server. I will ask for Your suggestions for the migration on another
> post soon).
>
> I have no much data to give. The backup log seems to be ok. The restore log
> seems to be ok until the process hangs.

How long are you waiting?

What does the server show as activity for the FB Server process(s)? How much CPU time is in use? How about disk I/O?

Is there much memory paging taking place? How much RAM does the server have? How much is in use?


Activating/creating indexes can require a lot of in-memory sorting prior to writing the index data to disk. So, you may be confusing "working" as "hung"!


Sean