Re: On-Bar Archive instead of Backup (or how to fake it)

Hello Vilmundur,

One of the ADSM API functions is called "dsmBindMC". One of the arguments to
this function is a flag indicating whether to use the archive copygroup or the
backup copygroup. Since the application (in this case, ON-BAR) calls dsmBindMC
(specifying the archive or backup copygroup flag), the type of copygroup used
is out of ADSM's control. Thus ON-BAR would have to provide some sort of option
to use an archive copygroup.

In answer to your two questions:

1) All backup versions of a particular file or object are bound to the same
management class. You can not have one backup version bound to one management
class, and another backup version bound to a different management class. If you
back up a file using a different management class, then *all* existing backup
versions will be rebound to the same management class. For example, suppose I
back up DSM.OPT using management class STANDARD. Then I do another backup of
DSM.OPT using management class NEWCLASS. This new backup version will be bound
to NEWCLASS, and the prior backup version will be rebound to NEWCLASS. So
expiration processing is based on the management class that the backup versions
are currently bound to, in my example NEWCLASS.

2) Yes, using a different node name in a different policy domain (with
management classes and copygroups that have the appropriate version and
retention settings) would give you the desired effect.

Andy Raibeck
ADSM Level 2 Support
---------------------- Forwarded by Andrew Raibeck/San Jose/IBM on 06/25/97
08:36 PM ---------------------------
08:36 PM ---------------------------

        ADSM-L AT VM.MARIST DOT EDU
        06/25/97 08:17 PM
Please respond to ADSM-L AT VM.MARIST DOT EDU @ internet

To: ADSM-L AT VM.MARIST DOT EDU @ internet
cc:
Subject: On-Bar Archive instead of Backup (or how to fake it)

> From cbehrens AT VNET.IBM DOT COM on 25-06-97 01:40:05:
> >From:         Vilmundur Palmason <villi AT NYHERJI DOT IS>
> >Subject:      ON-Bar Archive instead of Backup
> >To:           ADSM-L AT VM.MARIST DOT EDU
> >
> >Hi,
> >
> >Can anyone tell me if On-Bar (the Informix Backup Utility) can use ADSM
> >Archive instead of ADSM Backup.
> >The Redbook on Using ADSM to back up databases says that by default it
uses
> >backup
>
> Vilmundur, I do not believe there is anyway to change which copygroup
> onbar is using.  Why do you want to use archive instead of backup?
> thanks, Cyndie Behrens (IBM San Jose)

Thanks for your reply.  Here is why I am asking:

I am working with a customer who is evaluating ADSM.  They run many
applications on many  servers, the most critical ones on AIX machines.
They want to use ADSM for their daily backups but they also want to
take a copy each month for permanent storage.

The ADSM Archive process seems to be well suited for that and much can be
accomplished with a normal file system archive.  However, some of their
applications run on Informix.  So the question is, how do they set up a
schedule where they back up their database every day with a backup
retention setting of X days while taking a copy (or Level-0 backup/
archive) every month that never expires ??

Can anyone tell me if one of the following would work:

1.  Using a different management class
Which Backup Copy Group controls the number of days before an extra backup
copy of a file expires: the one that is active when the EXPIRE INVENTORY
command is run or the one that was active when the file was backup up ?

If it is the latter, then they could switch to a different management class
when doing the archive.

2,  Using a different node name
Can they create a special nodename in a seperate policy domain and switch
to that nodename by using a different dsm.opt file when they want to
archive ?



Many thanks to all who reply,



Vilmundur Palmason

Systems Engineer

Nyherji HF

Reykjavik

Iceland