Re: 725CPW vs 1450 IBM i Cloud System -- MIDRANGE-L

I like to use WRKSHRPOOL first and make sure the pool the HTTP job are running is set:

"Defined" -Paging Option-- for the pool in question is *CALC and not *FIXED.
<help>
*CALC
The system dynamically adjusts the paging characteristics of the
storage pool for optimum performance.
</help>

*fixed has a place but not for you needs.

Bryan

VERNON HAMBERG Owner via MIDRANGE-L wrote on 4/29/2024 4:16 PM:

Hi Brad

Has become kind of G(r)eeky to me to, but I am sure that pool 2 is *BASE, which is shared by everything other than interactive and printing and the machine pool.

If you press F11 a couple times in WRKSYSSTS, you'll see this kind of thing -

System Pool Reserved Max
Pool Size (M) Size (M) Active Pool
1 3309.37 1713.18 +++++ *MACHINE
2 28923.06 5.23 7054 *BASE
3 162.37 .00 41 *INTERACT
4 81.18 .00 28 *SPOOL

So system pool #3 is a shared pool for interactive work.

But there's another pool #, a subsystem pool id - if you press F14 in WRKSYSSTS, you work with subsystems -
Total ---------Subsystem Pools------------
Opt Subsystem Storage (M) 1 2 3 4 5 6 7 8 9 10
QBATCH .00 2
QCMN .00 2
QCTL .00 2
QHTTPSVR .00 2
QINTER .00 2 3
QSERVER .00 2
QSPL .00 2 4
QSYSWRK .00 2
QUSRWRK .00 2
Q1ABRMNET .00 2

Routing entries in a subsystem description are used to say which subsystem pool is used - and the system pool here shows 2's, 3, and 4

What a tangled web, eh? And I know, you didn't really ask, did you? :)

Cheers
Vern

On Mon, 29 Apr, 2024 at 3:44 PM, Brad Stone <bvstone@xxxxxxxxx> wrote:

To: midrange systems technical discussion

Message . . . . : SYSTEM INFO -> SYSTEM SERIAL NUMBER: xxxxxxx . SYSTEM
TYPE-MODEL: 9009- 22A. PROCESSOR FEATURE CODE: EP19. PROCESSOR GROUP:
P10.
MAX PHYSICAL PROCS IN SYSTEM: 20. CONFIGURABLE PROCS IN SYSTEM: 20.

Cause . . . . . : No additional online help information is available.

Here's WRKSYSSTS:

% CPU used . . . . . . . : 10.6<http://10.6> System ASP . . . . . . . :
238.6<http://238.6> G
Elapsed time . . . . . . : 01:18:40 % system ASP used . . . :
33.3983<http://33.3983>
Jobs in system . . . . . : 369 Total aux stg . . . . . :
238.6<http://238.6> G
% perm addresses . . . . : .007 Current temporary used . :
11173 M
% temp addresses . . . . : .015 Peak temporary used . . :
11292 M

Sys Pool Reserved Max ------DB------- ----Non-DB-----

Pool Size M Size M Act Faults Pages Faults Pages

1 1309.55 664.82 +++++ .0 .0 .0 .0

2 11435.25 8.64 354 .0 .0 1.1 3.1

3 1431.99 .00 358 .0 .0 .0 .0

4 143.19<http://143.19> .00 5 .0 .0 .0 .0

QHTTPSVR is set up now to run in pool 2. This is mostly greek to me.

Here's the subsystem info:

Total -----------Subsystem Pools------------
Opt Subsystem Storage (M) 1 2 3 4 5 6 7 8 9 10
QBATCH .00 2
QCMN .00 2
QCTL .00 2
QHTTPSVR .00 2
QINTER .00 2 3
QSERVER .00 2
QSPL .00 2 4
QSYSWRK .00 2
QUSRWRK .00 2

On Mon, Apr 29, 2024 at 3:37 PM Bryan Dietz <bdietz400@xxxxxxxxx<mailto:bdietz400@xxxxxxxxx>> wrote:

I know DrF is most probably correct, but "call QLZARCAPI" would show
what the LPAR config "looks like" for processor setup.

Bryan

Larry "DrFranken" Bolhuis wrote on 4/29/2024 2:29 PM:

So 725 is almost certainly 0.05 of a CPU, probably a Power8 server as
the smallest Power9 I have would be 850 for 0.05 CPU.

Clearly that's small and lower than the recommended minimum for 7.5 for
example.

Now will bumping to 1450 help? Almost certainly. Even if you ARE
uncapped, it will help a bunch. Every time a system goes back to the
queue for more CPU, it gets in line and when it gets time in the
processor the amount of time permitted is relative to the base CPU
allocation. Thus at 0.05 the smallest amount is allocated and when that
times out the process goes back in line again. Even at 0.10 the
allocation is doubled so right there it helps a bunch.

- DrF

On 4/29/2024 10:57 AM, Brad Stone wrote:

I am using a cloud system and I was set up with the lowest CPW

available.

It works in most situations except it is a little slow for websites
with a
lot of server side includes. This means that when a page is loaded

there

are 10 or more requests made at the same time to fulfill those SSI
requests
and it seems to overwhelm the CPU. (Think of it as 10 or more people
requesting a CGI web page at the same time).

Would a bump up from 725 to 1450 make much of a difference? Or is there
any tuning one can do to help?

I am thinking about changing the way the pages are developed to not
use so
many SSIs.. but they're so nice. :)

Bradley V. Stone
www.bvstools.com
Native IBM i e-Mail solutions for Microsoft Office 365, Gmail, or any
Cloud
Provider!

--

-- .
Bryan
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxxxxxxxx<mailto:MIDRANGE-L@xxxxxxxxxxxxxxxxxx>
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxxxxxxxx<mailto:MIDRANGE-L-request@xxxxxxxxxxxxxxxxxx>
Before posting, please take a moment to review the archives
at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxxxxxxxxxx<mailto:support@xxxxxxxxxxxxxxxxxxxx> for any subscription related
questions.