Replication Question

[cavscout]

I have created a replica of a database that has linked tables in it. Some of these tables are in other databases and some are linked via ODBC to other databases.

I created the replica with little issue and when I sync with the Design Master the replication process appears to complete successfully.

The problem is when I try to open and use the replica off-line I get a message saying that:

"Q:\Database.mdb isn't a valid path. Make sure that the path name is spelled correctly and that you are connected to the server on which the file resides."

This seems bizzare because as a replica I thought it didn't need to be connected to the server. Did my create replica action fail or is there a step I missed or am I totally mis-understanding what a replica can do?

Any advice please?

Thanks in advance.

 

[jeremie_ingram]

It was my understanding that the replication was for the purpose of keeping one "MASTER" database for development, and have the replicant for distribution. The end users would have the replicant, and you as the designer keep the master. You can change/edit/improve/modify the master, and with certain tools you can automate the update process to the replicants. Before you ever start the replication process it is strongly recommended that you make a backup copy of the database. You cannot convert the database back from a master, however it is possible to change a replicant into a database.
Sounds like you were wanting to create a working model that could be contained in a single machine. If this is what you are looking to do, then copy over the databases that are refrenced (from the Q: drive) to the local drive, and then use the linked table manager to edit the refrences back to the table you copied to the local machine. If you are only needing it for development reasons, you can get rid of the majority of the data from these items, and if you only refrence the table, get rid of the other unneccessary database objects from them as well.

 

[cavscout]

So, if I follow:

The diagram attached would not work by only installing replicas of the design master.

I would need to install a replica of each of the upperlevel databases on the laptops as well?

Forgive me if I'm not grasping a concept that is simple. I just want to make sure this plan is not practical before I change course.

Thanks for the reply.

 

[boblarson]

If the replicated databases have live access to the databases they are linked to then it will work if you make the replicas from the master and they have the same pathway.

What seems to be happening in your case is that you have set the path not by UNC (Universal Naming Convention) and have used drive letters to do so (example: Q:\").

I currently have an Access database with links to another Access database located on a LAN in Portland, OR that I needed replicated to Sacramento, CA. But, since we are on a slow WAN, we couldn't have the linked tables within link to the other database in Portland. So, I have a replicated database in Sacramento linked to the tables in the replicated version (also in Sac) of the other one in Portland.

What has to happen in that case, is that the design master must have the ability to connect to the remote one that you want to link on the other remote replication. Then, you can link it up and then create a replica down there. Then, you can relink the design master to the other design master, here in Portland, and then it will work. The links do not change when the replicas are synchronized.

I hope that helps as well.

Also, If you need true disconnected access then you will probably have to go to MSDE and SQL Server.

 

[Keith P]

I think that you may have misunderstood the purpose of replication. It is the data that needs to be replicated so that it may be used on such as a laptop disconnected from its network, and when subsequently connect back to the network, synchronised with your network replica and design master to update any data changes (normally a two way update).

I have never found it necessary to replicate the front end. I issue a new front end if changes to functionality are required.
Keith

 

[cavscout]

Thanks for your quick responses. I'm sorry to have such a tough time with this concept but for some reason I just am not grasping how this works.

I have a database on my server that everyone is working in to track quotes to customers. The data is stored on the server in the Q:\Database.mdb. This is the level where I link to other databases

The sales force uses a front end client to interface with the data.

I need the laptops to be able to use this data off-line (doesn't have to be real time or live, all they will be doing is entering follow-up notes).

I thought what replication would allow is at sync time the replica would take a "snapshot" of the server data and store it on the laptop to be used off-line until the next time the laptop hooks up to the network to sync allowing them to work with the same frontend as they would in the office.

Have I clarified my situation or am I still missing the point?

Pardon me if I am making you repeat yourselves, just trying to understand the limitations and where I'm expecting too much.

Thanks for your help.

 

[Keith P]

You have a front end database in which your forms, reports, queries etc. reside and a back end database holding your tables.

Each PC including your laptops will have a copy of the front end. Your networked PCs, excluding the laptops, will link across your network to the tables on your back end database (Q:\Database.mdb).

The Q:\Database.mdb needs to be either the Design Master or a Full Replica. Create a replica on each of the laptops from Q:\Database.mdb. You should now be able to use your laptops in the field. From now on the laptop will only ever update its own copy (replica) of the data.

When you next synchronise, any updates made on the laptop will be applied to the Q:\Database.mdb. Any changes made to Q:\Database.mdb since you last synchronised, i.e. updated by your networked PCs or from changes made by other laptops that have already synchronised, will be applied to the Laptop.

Aside from the logic of the above processes you may need to re-link the tables in the replica to the front end on the laptop.

For the use you describe replication is well suited. I have found the replication process works extremely well. Search the Microsoft KB under Access for “Replication White Paper” There are a number for different Access versions.
Keith

 

[Rob.Mills]

From my experience in replication it's best to never let any frontend connect to the design master. I keep that on my computer only where I can make changes as needed. Do you have a replicated BE on each laptop to which the laptops FE connects? That's the structure you need so they can be used offline. The workstations that are always connected all the time can link directly to the replica on the server. Then you just need to setup code to have your BE on the laptops sync with the server replica.

 

[cavscout]

Thanks guys, I'll take this and chew on it for a while. I think I see what your saying.

Thanks for all the great feedback.