Working offline

[dmantilla]

I am currently using Oracle for all our projects. Sometimes I work from home and use our VPN to access my models but it gets very slow. Is there a way to work offline? I've been copying the model to an .EAP file using Tools > Data Management > Data Transfer but then I have trouble when other people make changes at the same time.

Has anybody faced a similar situation?

Thanks in advance for any ideas.

Daniel

[«Midnight»]

I've seen something recently that might be similar. The poster was having performance problems across a link from AUS to NZ with a SQL Server repository.

The final resolution was a change to EA that came through in build 791 or possibly 792. From the sound of the Sparx post regarding the issue it seemed like it could affect other repositories.

You should try upgrading to the current build if possible. If that does not work then this should probably be reported to Sparx as a bug. You can reach the correct page from the Sparx site or from EA via Help / On-line Features / Bug Report Page.

[dmantilla]

Thanks for the tip, I just upgraded to build 792 and I will give it a try tonight from home. I will let you all know how it performed.

Daniel

[Eve]

Try looking into version control intead of the data transfer.  This will allow you to work at home on your EAP file and safely copy it back to the shared repository when you're done.

I recommend looking at The Version Control Overview in the help and A Version Control Whitepaper

[dmantilla]

I understand that using Version Control replaces the database repository approach, am I right?

Therefore, is working offline one of the major benefits of using VC vs. a database?

Daniel

[Eve]

I understand that using Version Control replaces the database repository approach, am I right?

No, Version control can be used with the database repository.  It will stop other users from editing any package you have checked out.  It then will only overwrite the packages you have checked out, meaning it's safe to assume no one else has changed it.

The problem is how you're getting your 'home changes' back into the repository.  It is overwriting everything in the repository when you do a transfer.

You don't even need to use the version control (unless you need the locking to ensure that no-one else changes anything in any package you're working on at home.)  Use the data transfer in the same way as you have been. Then when you get back to work don't do a transfer back, instead export the changed packages to XMI and import them into the model.  This will only overwrite those packages.

Of course, working via VPN is an option.  792 fixed one problem there, but it will still be comparitivly slow.