[OpenAFS] Re: Backup hierarchies ?

[Martin Schulz]

"Matthew A. Bacchi" <mbacchi@btv.ibm.com> writes:

> > b) From reading the administration manual, it did not get clear to me,
> > how to define the dump hierarchy. If I want to set up a weekly
> > incremental backup schme and keep the tapes 7 weeks each, do I need to
> > define this hierarchy eight times? The example in the section "To
> > display the dump hierarchy" makes me think so, though I find this
> > rather irritating.
> >
> > My way of thinking rather suggests to apply the same dump hierarchy to
> > different tapes at different times, thus one hierarchy sould
> > suffice. Can anybody explain that?
> 
> Martin,
> You can do it by creating 8 one week dumplevels, 

This one I would find irritating. Why should I define new hierarchies,
if I decide to buy more tapes and keep them longer? Does not convince
me.

> or by creating 1
> eight week dumplevel.  

How that? I want a full dump at the beginning of each
week. Incremental for the other days.

> Or, you can can create 1 one week dumplevel,
> which will not expire.  

It is mandatory for these dumps not to expire? I would like them to
expire after 7 weeks....

> This can be exploited by keeping your dump
> information in the backup database for a period of 8 weeks.  Then, if
> you need to restore from a tape that is 6 weeks old, you still have
> that information in the backup db, and you can use the tape(s)
> required.

I'd like to keep that info until the tape is reused (the minimum
imposed by the 7 week expiry). Or until I manually remove it. 

I still do not understand in the AFS admin guide (page 236 here) what
the sunday3 and sunday4 dump hierarchies are for.  Aren't they
redundant to the sunday1 and sunday2 hierarchies?

Otherwise asked: Can I have several dumps (for several weeks) of the
same dump hierarchy in my backup database? 

Sorry for beeing pedantic here, but for backup issues, it is better to
understand what I am doing...  I am actually searching for a missing
link a the logical chain:

1) Volume sets define a collection of volumes and form a template of what
to dump.

2) A dump itself is (basically) the data (or part thereof) of that
volume set at a given time.

3) A dumpset is a completely unrelated collection of dumps stored on a
set of tapes. 

3) A dump hierarchy is a template of the relationships of a certain
subset of dumps. The admin guide talks about "logical structure" that
makes me think here more in term of "classes" than of "objects" (in OO
parlance).

4) I am missing something that relates the actual incarnations of the
full/incremental dumps ("objects") to each other. I suspect, this data
structure is held in the backup database somehow, but how exactly?*)


Yours,
        Martin Schulz


*) My question is about the abstractions used, not about the internal
working of the database... 
-- 
Martin Schulz                             schulz@iwrmm.math.uni-karlsruhe.de
Uni Karlsruhe, Institut f. wissenschaftliches Rechnen u. math. Modellbildung
Engesser Str. 6, 76128 Karlsruhe