Shannon,
The application storing documents on the System i is WAF/400 which stores
extensive index information about the stored documents, including the
optical platter identifier. We have been able to extract key index
information and then organize the document extraction so that it is
processed in sequence by platter. That part was actually quite easy.
This issue, however, is complicated by the fact that the customer has more
documents than their juke box has locations for platters. So, the
specific optical platters staged in the unit has to be reviewed and
changed for each batch to make sure that the needed platters are actually
in the unit when we run the extract.
Rich Loeber
Kisco Information Systems
http://www.kisco.com
--------------------------------------------------------------------------
Shannon ODonnell wrote:
That's interesting! How did you limit it to pulling by platter? Did you
calculate what was on a platter first? I thought that Optical drives, like
normal dasd, wrote to whatever portion of storage was available at the time.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rich Loeber
Sent: Thursday, February 21, 2008 1:13 PM
To: Midrange Systems Technical Discussion
Subject: Re: File w/Large Number of Members
Shannon,
Initially, it was taking forever and we were only extracting about 20
documents a minute. If you do the math, I think that would end up taking
years to get all 7 million documents. But, we have since re-engineered
the extraction process to do it in "platter" sequence that that has sped
up the extraction process considerably. But, you're right, the juke box
is the limiting factor in this.
Rich
--------------------------------------------------------------------------
Shannon ODonnell wrote:
That's a big job.
Are you finding that it's taking a really long time for the optical drive
to
simply read those images so that you can convert them?
I've worked on something similar with an optical drive and the slowest part
of the entire process was the mechanical "performance" of the optical drive
itself.
Which made sense as that was the whole point of the customer getting off of
the optical drive in the first place.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rich Loeber
Sent: Thursday, February 21, 2008 1:05 PM
To: Midrange Systems Technical Discussion
Subject: Re: File w/Large Number of Members
Shannon,
The application is an image extraction of documents stored on an optical
juke box attached to the customer System i. They are moving off of the
System i and need to convert all of their documents into PDF files.
Each
member in the file represents a scanned document from the juke box.
There
are more than 7 million documents to be extracted.
Thanks for the suggestions so far. Some of them may have applicability,
but I have to be careful about tinkering with the application in the
middle of the conversion project.
Rich Loeber
Kisco Information Systems
http://www.kisco.com
--------------------------------------------------------------------------
Shannon ODonnell wrote:
I'd be interested in hearing about what kind of application would need to
generate that many members.
That's a new one on me.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Rich Loeber
Sent: Thursday, February 21, 2008 12:44 PM
To: Midrange Systems Technical Discussion
Subject: File w/Large Number of Members
Hello list,
I'm working with an application that has a work file that builds up a
large number of members while the application is running. It takes
several hours to complete. I even bumped into IBM's restriction of no
more than 32,767 members in a physical file with this one.
I'm have a minor issue with deleting the work file when I'm all done.
When the file has 30,000+ members in it, it can take more than an hour
for a DLTF to run. Has anyone ever run across this before? Are there
any tricks out there to speed up the file deletion process?
Rich Loeber
Kisco Information Systems
http://www.kisco.com
midrange.com
