Alan, thanks for the quick response. Either I'm not understanding you or 
something is missing from my initial post. The datasets don't change, there is 
no guaranteed backup frequency, and the one and only backup disappears,  not to 
be replaced by anything. At that time, the LBACKUP changes to zero.

I appreciate your offer for a suggestion. Our intention is to keep the latest 
three backup versions around for 70 days after the dataset gets deleted (after 
35 days). Perhaps your suggestion will shed some light on my misunderstanding 
of the documentation.

Thanks, also for the patch, but I'd rather get the backups working as intended.


David G. Schlecht | Information Technology Professional
State of Nevada | Department of Administration | Enterprise IT Services
T:(775)684-4328 | F: (775) 684‐4324 | E:[email protected]


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Staller, Allan
Sent: Wednesday, July 03, 2013 12:34 PM
To: [email protected]
Subject: Re: Missing HSM backups

Looking at the below, the underlying process seems to be a weekly activity.

>From the below, a maximum of 3 backups will exist at any given time. When the 
>4th backup is created, the 1st backup is deleted, *regardless* of age (at the 
>next backup cycle).
If you were to look closer, I believe you would see the backups begin to go 
away at ~day 56. (35 + 21) 

Check the column headings in the ISPF df/SMS application for the relationships 
among the BACKUP ATTRIBUTES variables.

In addition, check the dfHSM DEFINE BACKUP to ensure that things are being 
backup up when you think they are.
The is a patch available to allow dfHSM to delete datasets with no backup. 
Check the DBU attribute in the DFHSM DIAGNOSIS or DFHSM CUSTOMIZATION manual. 
(I forget which).
>From my HSM parms...   PATCH .MCVT.+431 BITS(......1.)     /* BYPASS SMS DBU 
>REQUIREMENT  */

If the intention can be specified, I will be happy to suggest appropriate 
changes to the BACKUP ATTRIBUTES.

HTH,


<snip>
I have date-derived dataset names  that get backups once created but after ~50 
days the backups start to disappear. Doing a Listcat ,the problem datasets show 
LBAKUP of 0000.000.0000. Listcat also verifies that the datasets all have the 
correct classes and storage group. Since the backups disappear, HSM refuses to 
delete the expired datasets.

There are no references to the problem datasets in the HSM job logs other than 
an occasional PARTREL.

Anyone seen anything like this or have any suggestions on where to look next? 
The management class and storage group definitions follow:
Expire after Days Non-usage  . : 35
Expire after Date/Days . . . . : 35
Retention Limit  . . . . . . . : 35
Partial Release . . . . . . : YES
Migration Attributes
  Primary Days Non-usage  . : 3
  Level 1 Days Date/Days  . : 5
  Command or Auto Migrate . : COMMAND
Backup Attributes,
  Backup frequency  . . . . . . . . . . . : 0
  Number of backup versions . . . . . . . : 3
     (Data Set Exists),
  Number of backup versions . . . . . . . : 2
     (Data Set Deleted),
  Retain days only backup version . . . . : 70
     (Data Set Deleted),
  Retain days extra backup versions . . . : 35
  Admin or User Command Backup  . . . . . : BOTH
  Auto Backup . . . . . . . . . . . . . . : YES </snip>

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email to 
[email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to