Load Schema with Remap has Killed my DB

[ConvertFromOldNGs]

by Mike May >> Thu, 30 Oct 2003 12:12:35 GMT

I have two copies of Jade, one on my personal PC, the other on a development server. I created a small test schema & app on my PC and needed to extract and load it onto the development DB. My personal Jade install folder is D:\Jade60; the map file for my app points to D:\Jade60\system\MMTemp.dat. Jade on the dev server is installed in E:\JadeDev01. So, I created a remap table for my schema that remapped to D:\JadeDev01\system. However, the load failed with error 5027 (dir name invalid). The logs did not tell me what the nasty dir name was, but I assume it was because D:\Jade60\system does not exist on the dev server and the load would have reied to find and rename the existing map file. I can no longer open the Jade database on the server and would welcome advice as to how I can get it back.

[ConvertFromOldNGs]

by Mike May >> Thu, 30 Oct 2003 15:45:27 GMT

Sorry, don't know why this one got in twice. I've sorted out my DB - it wouldn't open in single-user mode but the RNA worked fine so I was able to connect from Pres Mgr and delete the offending schema.

[ConvertFromOldNGs]

by BeeJay >> Fri, 31 Oct 2003 8:17:02 GMT

I suspect that the reason why you couldn't open in singleUser is because you were using the same usercode as you used when loading the schema and it was therefore trying to access the non-existent mapfile on the non-existent drive to set your application.

You may find it easier rather than using a remap table to simply remove the path from your mapfile definition. Then, the mapfile will go into the same directory as you specified on the path= statement on your shortcut and will work on both your d: drive and your e: drive installations of JADE without any extra effort required. You'll also have the added advantage of being able to have this same schema loaded onto two different installations of JADE on the one machine whereas with a path defined for your mapfile you will run into problems trying to run two copies of this schema on a single machine.

Refer to FAQ 17673 on this Web Site which discusses this and has sample code for removing a hard-wired mapfile - although you'll need to adjust the sample method logic slightly as the DbFile::_rename method now has a single string parameter for the path.

ie: dbf._rename( null );

Cheers,
BeeeJay.