Hi,

I am not experienced in using hpl and I need to know how to configure and
use
the HPL to reorg large tables.
We have to reorg some very large tables in just 5 hours (24*7). So I cannot
use
SAP's sapdba. Somebody out there who has real advise for me ?
Thanks in advance !!

Regards

Torsten Barth
---------------------------------------------------------------------------------------------------------

Barth Rechenzentrum + Systemhaus GmbH & Co. KG
Organisation & SW-Entwicklung
Mühleneschweg 5
D- 49090 Osnabrück
Telefon +49 541 609 10 - 88
Telefax +49 541 609 10 - 89
mailto: tbarth@barth-rz.de


sending to informix-list

Torsten Barth wrote:
>
> Hi,
>
> I am not experienced in using hpl and I need to know how to configure and
> use
> the HPL to reorg large tables.
> We have to reorg some very large tables in just 5 hours (24*7). So I cannot
> use
> SAP's sapdba. Somebody out there who has real advise for me ?
> Thanks in advance !!
>

Please give IDS version & the platform you are on.

You should read about the onpladm utility, if it is
available in your version:
http://www-106.ibm.com/developerwork...07carlson.html

Also search the library for the load articles written by J. Parker
(find pointers in c.d.i. using the search terms +parker +load + faq)

Be aware that I have *absolutely no idea* what you must do with SAP
while
unloading & table drop & reloading & & index recretate &
index rebuild & !!update statistics!!

dic_k
--
Richard Kofler
SOLID STATE EDV
Dienstleistungen GmbH
Vienna/Austria/Europe

Hello Torsten,

you can use ipload (GUI to HPL) to configure the lot easier when newbie;
you need X though.

Done this stuff in the past; table without blobs are very good candidates.

used fifos to do the copying see mkfifo.
(CAREFULL PLEASE IF os mucks up you're in trouble)

build 2 jobs one for load and one for unload in express mode please.
Afterwards create level 0 archive.

the unload writes to the fifo(s) and the load job reads from it.

this means you build a new table to store the data in; later when reloaded
drop the old table and make sure things like views etc are also corrected.

See you

Superboer.










Torsten Barth <tbarth@barth-rz.de> wrote in message news:<cdipar$6sf$1@news.xmission.com>...
> Hi,
>
> I am not experienced in using hpl and I need to know how to configure and
> use
> the HPL to reorg large tables.
> We have to reorg some very large tables in just 5 hours (24*7). So I cannot
> use
> SAP's sapdba. Somebody out there who has real advise for me ?
> Thanks in advance !!
>
> Regards
>
> Torsten Barth
> ---------------------------------------------------------------------------------------------------------
>
> Barth Rechenzentrum + Systemhaus GmbH & Co. KG
> Organisation & SW-Entwicklung
> Mühleneschweg 5
> D- 49090 Osnabrück
> Telefon +49 541 609 10 - 88
> Telefax +49 541 609 10 - 89
> mailto: tbarth@barth-rz.de
>
>
> sending to informix-list

Torsten,

Not sure about SAP but I used onpladm before (Probably with IDS 9.21 &
IDS 9.30). Have a look at example/syntax below which I used:
--------------------------------------------------------------------------
onpladm create project <project_name>
onpladm create job <jobname> -n -p <project_name> -d <device_name> -D
<database_name> -t <tabel_name>

Example:

onpladm create project hari
onpladm create job table1 -n -p hari -d table1.hpl -D db1 -t
owner.table1

Also check the syntax below:

usage: onpladm create job <jobname> [-p <project>] -d <device> -D
<database>
-t <table> [-f[luadpcN]] [-z <formattype> | -n] [-T
<targetserver>]
[-S <onploadserver>] [-M <devicesize>] [-B <blocksize>]
where,
-p project name
-d device name, further qualified by flags
-D database name
-t table name
-f flags,
l means load job, default will be both load and
unload
u means unload job
a means device is a device array, default will
be file
d means device is a tape
p means device is a pipe command
c means runmode DELUXE with replication, default
will be
EXPRESS runmode
N means runmode DELUXE without replication,
default will
be EXPRESS runmode
-z type of format to be used, default is Delimited type
D means DELIMITED_FORMAT
FI means FAST_FORMAT
FA means FIXED_ASCII_FORMAT
FB means FIXED_BINARY_FORMAT
C means COBOL_FORMAT
CB means COBOL_BYTE_FORMAT
-n no-conversion fast-job
-T Target server, where load/unload is to be done
-S Onpload server containing the onpload database
-M device size, if device is a tape
-B block size, if device is a tape
--------------------------------------------------------------------------