Page 1 of 1

DUMP and RESTORE in a PRODUCTION

PostPosted: Tue Mar 20, 2012 5:31 pm
by samb01
Hello,

i have to instal a politic of DUMP and RESTORE in a classi production.

I juste want to know how do you do in your site.

For example :

1) Dou you DUMP the dataset at the end of a the TWS application ?

2) Whith ADRDSSU ?

3) When do you delete the dataset ? At the beginning of the TWS application ot at the end with the DUMP/DELETE of ADRDSSU ?

4) How do you restore your dataset ? How do you find quicly the number of the tape to restore ?


Thank's for your answer.

Re: DUMP and RESTORE in a PRODUCTION

PostPosted: Tue Mar 20, 2012 7:27 pm
by steve-myers
samb01 wrote:Hello,

i have to instal a politic of DUMP and RESTORE in a classi production.

I juste want to know how do you do in your site.

For example :

1) Dou you DUMP the dataset at the end of a the TWS application ?

2) Whith ADRDSSU ?

3) When do you delete the dataset ? At the beginning of the TWS application ot at the end with the DUMP/DELETE of ADRDSSU ?

4) How do you restore your dataset ? How do you find quicly the number of the tape to restore ?


Thank's for your answer.
First, "dump" in this context is not the correct word. "Backup" would be a better choice. One seldom "dumps" a single dataset, In this context "dump" applies to multiple, usually unrelated datasets.

Second, "TWS application" is basically meaningless. TWS is a job scheduling system.

Third, using ADRDSSU to backup a dataset (or more likely a group of related datasets) into its "dump" format is certainly possible, but not advised.

Fourth, you do not delete a dataset before you back it up. The only common exception to this rule is many JCL defined jobs conditionally delete a dataset and then reallocate it, with the implied context that there is already a viable backup of the dataset.

Fifth, when you backup a dataset, and the tool used to perform the backup is immaterial, you are creating a new dataset. You catalog this new dataset, so if you need to restore the dataset contained in the backup dataset you refer to the backup dataset as a cataloged dataset in the usual way.

Re: DUMP and RESTORE in a PRODUCTION

PostPosted: Tue Mar 20, 2012 8:28 pm
by samb01
Heelo steve-myers and thanks for your answer.

Sorry, my english is not very good...

I anderstood your answer except the third...

I think everybody use ADRDSSU to dump file

//SAVE     EXEC PGM=ADRDSSU                                         
//SYSPRINT DD SYSOUT=*                                               
//OUT      DD UNIT=VTS2,DISP=(,CATLG,KEEP),VOL=(,,,10),             
//         DSN=RDT.FCVT.DUMP                             
//DUMMY    DD DUMMY                                                 
//SYSIN    DD *
DUMP OUTDDNAME(OUT) OPTIMIZE(3) FORCECP(0) TOL(ENQF) -       
     DATASET(INCL(RDT.FCVT..**,A.UTRT12.CVFT.**))
IF LASTCC LE 04 -                                           
THEN DUMP OUTDDNAME(DUMMY) OPTIMIZE(3) FORCECP(0) -         
     DATASET(INCL(RDT.FCVT..*)),DELETE             
                           



Re: DUMP and RESTORE in a PRODUCTION

PostPosted: Wed Mar 21, 2012 6:37 am
by steve-myers
As you may have deduced, I am not a big fan of ADRDSSU.

The reason I do not like it for the purpose of performing a backup of multiple datasets into its large container datasets is constructing the commands to restore a dataset from a container datasets is difficult (for me, anyway), to say the least.

If I understand your job, you're telling ADRDSSU to read the source data twice, once for the real dump, and a second time, for the "dump" used to delete the source data.

Since you're deleting the data, why are you performing this backup?

Re: DUMP and RESTORE in a PRODUCTION

PostPosted: Wed Mar 21, 2012 6:58 am
by Robert Sample
Different people, different experiences. I've had the occasional problem getting the command right in ADRDSSU, but once the command is right, I've never had any problems restoring data using it. Our site uses it for all of our backups, including disaster recovery, and even the disaster recovery restores go smoothly.