Re: First time buying System i5

[Tom Liotta <qsrvbas@xxxxxxxxxxxx>]

Glenn:

An additional quick note...

We have a _lot_ of source members and quite a few source files. With 
developers coming and going, learning our standards and the CMS we 
use, an odd source member might get promoted to the wrong location 
or simply lost in one way or another. Rare, but it's happened.

A few years back, I created an old CL program that:

1.  Did DSPOBJD of all physical files on the system to an *outfile.
2.  Ran a simple SQL DELETE to delete all non-source files from there.
3.  Then RCVF'd each record to do DSPFD TYPE(*MBR) OUTMBR(*N *ADD) 
to a second *outfile.

That would give me a file listing all source files on the system and 
another file listing every member in every source file on the 
system. And that would let me run SQL SELECTs against member names 
to locate stray variations of source members.

Trivial tool, but extremely handy when it's needed.

Previously, I would only schedule it to run over infrequent 
weekends. It didn't really need constant updating since it was just 
for my own use. But now, I schedule it for every night and even run 
it on demand whenever I feel like it.

Today, I ran it in the middle of the day just to see how it ran --

10989 total physical files
4123 source files after non-source were deleted
50259 source members in the resulting file

Wall-clock times: Start 11:14:11.198352, End 11:25:29.703472

It's about as inefficient a way to do it as I'd ever program for; 
but just over 11 minutes of wall-clock time in the middle of our 
day... _plenty_ fast enough for a developer (IMO). And I suspect 
that all three drives in the RAID set were exercised a little.

Tom Liotta

Glenn Ericson wrote:
> Tom ,
>   so we can put this in proper perspective,  what prior systems has 
> Power Tech placed in each LPAR?
> Not sure  what  throughput stress  testing  application for  security
> 3 drive  RAID sets might be  fine for lite workloads. That 4th drive 
> makes tons of  difference.