about Datastage Job Backup

mandyli · Post by **mandyli** » Thu Apr 13, 2006 6:17 am

Hi

I want to take every week backup of my current datastage project. currently i am exporting my whole project as .dsx file. any other way in DS 7X automatically taking datastage backup.

please through some inputs.

Thanks
man

chulett · Post by **chulett** » Thu Apr 13, 2006 6:48 am

That is the 'best' and most common way to backup jobs. The other way is a 'normal' filesystem level backup so that the full DS installation and things like your hashed files are backed up as well. Also you'll find things like 'uvbackup' (?) to backup individual hashed files, but I don't think many people (other than the Grognards) know about it or use it.

I export all of my development projects every day using a script someone posted over on ADN.

ray.wurlod · Post by **ray.wurlod** » Thu Apr 13, 2006 3:24 pm

My automatic export strategy is a BAT file that executes dscmdexport (after renaming yesterday's export after deleting the previous renamed export file). If you don't mind embedding a clear password (and relying therefore on Windows security to prevent others from seeing it), you can schedule the BAT file using AT.

DeepakCorning · Post by **DeepakCorning** » Tue Jan 23, 2007 3:19 pm

Chullete ,

When you say "The other way is a 'normal' filesystem level backup so that the full DS installation and things like your hashed files are backed up as well" do you mean a simple copy paste in come other place?? Will this make sure that the hash file are intact? Also if I create a new project with a different name and paste all these files in it then will the Hashed file will be in sync or it does not work liek that?

I am really sorry to bring up such an old topic...

satisht · Post by **satisht** » Sun Jan 28, 2007 12:09 pm

the method you are using is very easy and mostly i am doing in the same way

ArndW · Post by **ArndW** » Sun Jan 28, 2007 3:49 pm

DeepakCorning wrote:.....simple copy paste in come other place??... Also if I create a new project with a different name and paste all these...

A UNIX or Windows copy-paste of a DataStage account will not work. The engine stores information about secondary indices inside hashed files using an absolute path, so any changes to copies of indexed hashed files will update the original index, making it incorrect for both accounts! DO NOT use an OS level copy in this manner unless you know which files are indexed and know how to correct the paths (this in not documented as part of DataStage).

ray.wurlod · Post by **ray.wurlod** » Sun Jan 28, 2007 6:47 pm

And it's not just indexes. A DataStage project is a schema, and there are system table entries that need to be made/corrected also.