Page 1 of 1

Backup HashFile

Posted: Tue May 18, 2004 9:02 am
by zhecov
Hi ...

knows somebody of some tool to make backup of hash file or some form of make.. thanks

Re: Backup HashFile

Posted: Tue May 18, 2004 1:38 pm
by ogmios
Copy the physical files of the hashfile when no DataStage jobs are running on it.

Anyway you should be able to recreate hash files at will: I've had numerous hash files go corrupt on me. So if the only place where you store your data is a hash file I would rethink your design.

Ogmios

Posted: Tue May 18, 2004 4:27 pm
by ray.wurlod
A job that populates a hashed file can take a backup before doing so. On the inputs link to a Hashed File stage there is a check box called "Backup existing file" - click on the Help button for more information.

As other posters have noted, because a hashed file exists in the operating system, you can copy the operating system objects (a dynamic hashed file is a directory containing three files DATA.30, OVER.30 and .Type30). It is vital that nothing be using the hashed file when it is backed up. Metadata for a hashed file live in a separate file, called D_hashedfilename.

There is a backup utility in DataStage, called uvbackup (and uvrestore for restoring from a uvbackup image). The same caveats about "file in use" apply if you choose to prefer this utility.

Re: Backup HashFile

Posted: Wed May 19, 2004 3:17 am
by arunverma
Hi ,

While creating hash file , you can also give path , where you want to store ( Application working Directory ) then you can take backup , If you are not chhosing path then hash file create default directory on ETL server .

Arun

Posted: Fri May 21, 2004 2:49 pm
by zhecov
ray.wurlod wrote:A job that populates a hashed file can take a backup before doing so. On the inputs link to a Hashed File stage there is a check box called "Backup existing file" - click on the Help button for more information.

As other posters have noted, because a hashed file exists in the operating system, you can copy the operating system objects (a dynamic hashed file is a directory containing three files DATA.30, OVER.30 and .Type30). It is vital that nothing be using the hashed file when it is backed up. Metadata for a hashed file live in a separate file, called D_hashedfilename.

There is a backup utility in DataStage, called uvbackup (and uvrestore for restoring from a uvbackup image). The same caveats about "file in use" apply if you choose to prefer this utility.
thank's