Duplicate database

[mf]

Hello,

I am just upgrading our server from version 8 to 11. I did not know until now that backups were now each in their own folder. I understand the upside of allowing multiple backups and preventing them being overwritten.

Having a fixed location was a plus in this situation: we need a duplicate of the database at a remote location - the data is replaced every night and using "day-old" data is not an issue and performance is better than using the network. So the script used by the remote machine was "pulling" a backup of the database from the same location every night. Now, with a new name every night, it no longer works...

We can't do a file copy of the live data, can't "save as" while the files are hosted...

My question: is there a simple way to duplicate/backup a database in a specific location? thanks for your thoughts.

Edited June 29, 2010 by Guest

[Steven H. Blackwell]

So the script used by the remote machine was "pulling" a backup of the database from the same location every night. Now, with a new name every night, it no longer works...

This is essentially backwards of what should have been done. And it poses risks to the server and the files.

[color:red]Push the files out to the remote location instead. See the backups information from Wim Decorte's website for several approaches to doing this.

Steven

[mf]

Whether push or pull, I was looking for a way to avoid a folder name that changes with every backup... Or a method to obtain a copy/duplicate of the database in a fixed location.

[VolkerKrambrich]

What you want to do:

Inside the standard folder for FileMaker Server backups, do create a folder and name any way you like it - "MyCopyForRemoteSite". Now create a schedule in FMS Admin Console to save your database file(s) once a day into that folder, keeping only one version (meaning overwriting the contents tomorrow.)

Now you have the folder where to look for data from you system script; that script will first prepare, then push (copy) the data to the remote site.

In my scripts. I first copy data to my work directory, then use zip with the options -rj (r traverse directories recursively, j junk, no directories included) and *.fp7 for the file mask. Voila, in the resulting archive.zip are only the files included; no timestamped folders.

Make sure to rename the archive.zip or do whatever find appropriate to identify todays' copy from yesterdays'.

Regards

Volker

[mf]

I never thought of using these parameters with winzip. I will give it a try, thanks!