Re: Reducing ODBC connection impact

[DrFranken <midrange@xxxxxxxxxxxx>]

So this had me laughing out loud.  I just cranked up an old 170 to 
assist a customer migrating S/36 programs, OCL, and data to IBM i. As 
they only had ancient QIC I needed the old girl to read that tape.

I was SHOCKED at how slow it was and I remember installing that very 
machine brand new 'some years ago' and thinking how impressively fast it 
was back then! Just restoring the S/36 source and OCL (about a dozen 
libraries) took overnight. Including the customer's data the *SAVF I 
created for them was a gargantuan ginourmous hugemoungous 375MB. Yes MB.

So while I do suspect there are some things you can do to assist a bit 
in lowering the impact of these inserts, such as lowering the priority 
in the COSD as others have suggested, I'm glad it's a hobby machine and 
not running real production!!

        - Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 12/6/2019 7:07 AM, Patrik Schindler wrote:
> Hello,
>
> I'm doing some data transfers between my 150 (running V4R5) and a Linux VM elsewhere via ODBC. Especially INSERTs are very costly and slow down other (interactive) tasks on the box considerably. For that reason I want to lessen impact of the QZDASOINIT PJs in the QSERVER Subsystem. Also I'd prefer to do that in a way that the setting(s) survive an IPL.
>
> I still didn't get a solid grip about jobs and work management. I have a rough idea about relative job priorities and time slices. My understanding of subsystems is even more terse. How would you achieve the described goal? No detailed steps necessary. Primarily I want to collect ideas how to do it (properly).
>
> Please refrain from comments like "get a current machine" or likewise. I'm a hobbyist in Midrange terms with a tendency to retrocomputing. Thank you. :-)
>
> :wq! PoC
>
> PGP-Key: DDD3 4ABF 6413 38DE - https://www.pocnet.net/poc-key.asc