Doing a partial upgrade 5.0 to 6.0 to use Web service?

lagerberg

Hi,
I have a client running Nav a 5.0 sp1 database with Nav2009sp1 clients (Classic), where we would like to use just the web service functionality to run a codeunit.
I have tried installing the service tier and web service, and they do start, but (as expected) are missing things in the database.
From reading the forum I understand there are a couple of things that are needed, and the first thing I am about to do is switch on the compilation of objects.
I understand that is done with the switch "Enable for Microsoft Dynamics NAV Server" in database settings.
Now, my question: Is it safe (and maybe reversible) to do this? I am working in a production environment and cannot risk stopping the classic environment from working.

Cheers

matttrax

Why on earth would you be doing this in a production environment and not a test database?

lagerberg

Hi, thanks for the near instantaneous reply!
Ok, so thats the answer then.. It's not safe?
The reason I would do it if it was safe is really a matter of time and cost... there is no testing environment at this (small) customers site.
Cheers

dmccrae

Enabling this setting will generate metadata in a system table necesary for using the NAV Server. This in itself will not prevent use of the classic client as a 2-tier client, and when disabled it will simply delete this additional metadata. So yes it is reversible. Take note of the following though:

1. Changing this setting may require that you first put the database into single-user mode, then it takes quite a long time to generate the metadata (dependent on the number of objects in your database, it is typically around 20 mins). During this period, other clients will not be able to connect because of the single user restriction.

2. The additional metadata takes up some space in the database so there should be space available to do this.

For these reasons a quick test on a different database might be a good idea.

matttrax

I'm not saying it isn't safe, it just makes more sense to do it in a test database first. Every customer / database is a little different and you never know what might blow something up. They should have a test database anyway for deploying new modifications. As for the time and cost, surely it can't take that long to backup a database (or hopefully just take last night's backup) and restore it.

In my opinion, there is hardly anything that it is worth risking bringing a production database down for.