Not yet, Phil. I'm gonna collate some data from the [operation_messages] and see if that points me in a direction.

It shows the "For Loop Container" task.

The variables in each stack are unique to that stack (they're all named for the stack in which they exist).

This is failing when the package runs from a SQL Agent job.

I forgot to mention that *most of the time*, it works. We just get this failure occassionally.

Yeah I've already implemented a workaround in this case but wanted to know for future implementations where this might not be possible.

Grrr... Such a simple answer! For some reason I'd set it to 0.

Many thanks.

Thanks for that. There was a slight issue in that the day portion being returned was 2nd of the month for a date duration of 1 day - because 0...

Thanks for the feedback. Yes, the initial restore was greatly slower as the files were initialised but the regular refresh is obviously using pre-existing files so there's no speed improvements...

Thanks Grant. Yeah, that's what I thought re: a restoring DB.

I'll see if there are any infrastructure-level options available to us.

Thanks again.

Note: I've managed to make this work by changing "pattern2" to:

pattern2 = "[\x22\x27\x2c\x3f\x5c\x5c\x60]"

but I don't know why I need to escape the backslash in the pattern2 value...

Found the answer. On the secondary server run:

SELECT [secondary_id] FROM [msdb].[dbo].[log_shipping_monitor_secondary] WHERE [secondary_server] = N'SECONDARY' AND [secondary_database] = N'DBNAME'

Turns out that the timeout setting for the data driven subscription is 30 seconds (default) and this is where the issue was occurring. Didn't realise that this is what was...

Alex. I had that thought but wasn’t sure if that would differentiate between the database being encrypted and the backup of the database being encrypted. Peter

RESOLVED:

It turned out that the availability group was the issue. Looking at sys.databases I saw that the log_reuse_desc was AVAILABILITY GROUP so it was replication that was preventing...

Hmm... Changing this through the normal services interface worked. When I set the account to use I got a message stating that the "log on as a service" right was...