Hello everyone,
I want to migrate all hash file from one server to another, does anyone know how to do it? I tried to physically move hash files but the ETL job on the new server can not recognize it.
this is the error message:
J_HASH_xxxxx..Trans_Dummy: |J_HASH_xxxxx..HASHxxxxxx.Trans_Dummy: DSD.UVOpen "HASH_xxxxx" is already in your VOC file as a file definition record.
moving hash files
Moderators: chulett, rschirm, roy
There are two ways that hashed file stages access the files - using an explicit path and using the (default) project directory. Hashed files that are referenced by absolute or relative UNIX/Windows paths can be copied from one host to another (being careful to do so when the files aren't in use and in preserving any hidden files and access rights) and will not cause any problems.
The second type is not as easy to migrate or copy since these files also create a file descriptor record in an internal hashed file (called the VOC for Vocabulary) in the project. The error message you posted is means that DataStage is trying to create this file although it already exists as a pointer in the VOC, even though the OS file it points to might not exist.
You should run or verify your job on the new machine to let DataStage create the hashed file pointer and the physical hashed file first. Once that is done you can copy and overwrite the file in the project directory with the original one.
The second type is not as easy to migrate or copy since these files also create a file descriptor record in an internal hashed file (called the VOC for Vocabulary) in the project. The error message you posted is means that DataStage is trying to create this file although it already exists as a pointer in the VOC, even though the OS file it points to might not exist.
You should run or verify your job on the new machine to let DataStage create the hashed file pointer and the physical hashed file first. Once that is done you can copy and overwrite the file in the project directory with the original one.
</a>