EMC NetWorker discussion <NETWORKER AT LISTSERV.TEMPLE DOT EDU> wrote on 
09/04/2008 11:54:04 AM:

> I'd be interested in knowing the final verdict from support. Under
> normal circumstances, how would you ever know the actual order of a
> multi-tape to scan? 

You scan each volume twice, apparently. See below.
> 
> This sounds like another huge flaw in Legato logic to be aware of,
> unless there is an official explanation and reason for having to do this
> the way you did.

This is what Tech Support told me:

Please review and follow the steps on page 310 of NW 7.4 Multiplatform 
admin's guide below for the details. 

Which says:

To rebuild the save set?s entry in the client file index and the media 
database:
1. Log in as root or Windows Administrator.
2. At the command prompt, run the scanner program on the volumes that 
contains
the appropriate file or files:
scanner device_name
3. Use the output from the scanner program to determine:
? Whether the save set to be rebuilt is on this volume.
? Whether to reintroduce the contents of this volume into the online 
indexes.
(Locate all the volumes that contain this save set ID.)
4. If the save set is found on multiple volumes and the order in which the 
volumes
were written is unknown, complete this step. Otherwise, skip to Step 5.
a. On each volume that will be reintroduced into the online indexes, run 
this
command:
scanner -m device_name
The media database is updated with information from each volume.
Note: If the volume contains data from an earlier version of NetWorker, 
there may be
no pool information on the volume. In this case, the volume is considered 
to belong to
the Default pool. To assign the volume to another pool, use the -b 
pool_name option in
this step. If the volume already belongs to a pool, the -b option will 
have no effect.
b. Query the media database to determine the sequence in which the volumes
were written:
mminfo -a -v
In Step 5, the volumes must be reintroduced into the online indexes in the
same order in which they were written.
5. Starting with the first volume that was written, run the scanner 
program:
scanner -i device_name
The scanner program prompts for a new volume until you terminate it. Scan 
in
the remaining volumes in the order in which they were written.

IOW ... scan *twice*. Scan each tape once with -m, to put it in the media 
database. Then query for the sequence of the spanning datasets, using 
mminfo. Then scan *AGAIN* with -i, in the sequence order.

Which strikes me as an extremely BRAIN DEAD way of doing it ... if you can 
figure out the sequence by querying the media database, why can't NW do 
the same thing? Why should I have to scan every tape *TWICE*, and in a 
specific order at that? If I can tell by looking at the fragflags where 
the head, middle, and tail of the saveset is, and which volume it is on, 
how can NW not do the same? It seems almost inconceivable ...

To me, that's just really bad design or programming. Very user unfriendly, 
and very time-consuming. I'll never know if the old saveset I wants spans 
multiple tapes. So I'll have to scan each old volume twice, *just in case* 
the saveset spans multiple tapes.

Well, I'm extremly disatisfied with THAT behavior .... the only way around 
it that I can see is extraordinarily long browse/retention times. Right 
now, I have 2 months browse, and 2 months retention. I'll have to increase 
that to a year or two, otherwise every recover I want that is older than 2 
months will need to be scanned twice.

At the moment, I am one unhappy customer ....

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER