Batch Processing

        Hi ADSM-Folks (User Community)
        and IBM-People (Development & Support),

I am using ADSM since it was a baby (beta4 - more than two years ago),
but still there are "features" not fixed within that long period of time
despite various DCRs were sent to our TAs and SEs.

I'm talking about batch processing using B/A- and Admin-CLI-commands.
Currently, we want to do some accounting on a per filespace basis using
dsmadmc's "query occupancy" command. Unfortunately, "q occ" causes some
trouble.

As with many query commands output is mostly formatted in a way not very
suitable for batch (ie. script or other automatic) processing. Even with
format=detailed (if applicable at all, not with q occ) some hyphenation
is done by ADSM.

And there are cases where it is really impossible to extract the original
information out of ADSM's formatted output (see example of a PC's filespace
name containing blanks below: the name of the filespace might be - at a
first glance - "C650 user_disk" or  "C650    user_disk" or similar).

This is very annoying as there is no means to circumvent this unnecessary
behavior. There should be flags like -format=raw to force record-type
output together with -fieldsep="<;, or what you like>" to specify a non-
default field separator - and this kind of problems would be solved once
and forever. (Will IBM ADSM Development take over this request? You should
seriously think about it.)


In addition the type of output of an interactive session depends on the
current environment, eg. window width (!). For example in a wide window
(and batch environment as well) you might get these output lines from
"q occ ..."

Node Name                       Filespace    Storage      Number of       Space
                                Name         Pool Name        Files    Occupied
                                                                           (MB)
------------------------------  -----------  -----------  ---------  ----------
MACCENTRIS.PE.BA.DLR.DE         C650         PE_3590B         9,034    2,688.55
MACCENTRIS.PE.BA.DLR.DE         C650         PE_3590B         9,034    2,688.55
                                 user_disk

whereas in a smaller window dsmadmc will output the following:

          Node Name: MACCENTRIS.PE.BA.DLR.DE
     Filespace Name: C650 user_disk
  Storage Pool Name: PE_3590B
    Number of Files: 9,034
Space Occupied (MB): 2,688.55


Is there any chance to force the latter type of output in a batch environment?
Even redirecting or piping output within a narrow window leads to this
unwanted broad-type output. So dsmadmc decides which type of output to use.
How? Can we overwrite it?

We do not want to switch to APIs just because of this "problems".
I would truly appreciate any helpful comments on that topic.

        Rainer


 ----------------------------------------------------------
 Rainer Strunz,   DLR    German Aero Research Establishment

 Rainer.Strunz AT dlr DOT de    Research Centre Oberpfaffenhofen
 Tel +49-8153-28-2933    Dept. WT-DV-ITS
 Fax +49-8153-28-1136    D-82234 Wessling
 ----------------------------------------------------------