Problem solved - Many Thanks to all who responded.
Mark, David - My initial dismay at how file protection is handled in RACF
turned out to be unfounded (of course!).
I was primarily thinking about the number of user or group profiles I'd
have to define in order to adequately protect my system.
It
Are there any hints/tips to improve the performance of an ISPF application
that uses BPXWUNIX? In running on my normal system the performance is very
slow compared to a friends system running on a zPDT.
I couldn't find a SHARE presentation (but I may not have used the correct
keywords) and the
Lionel,
The secret is that zPDT systems are not heavily loaded. As you load the
system (or start mot guests, if running under z/vm), the system starts to
slow down. BPXWUNIX starts a new address space, so by definition it is more
cpu consuming, that affects elapse.
ITschak
On Mon, Nov 4, 2019 at
That was my fear 😊
I did find these which I implemented in my .profile and I'm not seeing any
visual improvements:
export _BPX_SHAREAS=YES
export _BPX_SPAWN_SCRIPT=YES
Lionel B. Dyck <
Website: http://www.lbdsoftware.com
"Worry more about your character than your reputation. Chara
That's because _BPX_SHAREDAS is related to a main shell program and its
forks. not to the Rexx (or whatever the caller is) and the unix shell.
On Mon, Nov 4, 2019 at 1:29 PM Lionel B Dyck wrote:
> That was my fear 😊
>
> I did find these which I implemented in my .profile and I'm not seeing
> a
But when rexx, under ispf, invokes bpxwunix won't the sharedas help to prevent
the fork?
Lionel B. Dyck <
Website: http://www.lbdsoftware.com
"Worry more about your character than your reputation. Character is what you
are, reputation merely what others think you are." - John Wooden
-Ori
No. the fork is a unix issue. the first shell task will be started as a
separate asid and then, if _BPXSHAREDAS is on, called shell programs will
run in the same (new) asid. You can see in the syslog the evidence of the
new ASID started. It is usually has the id of the caller plus numeric
value.
Thanks for the education.
Guess I need to find a faster system 😊
Lionel B. Dyck <
Website: http://www.lbdsoftware.com
"Worry more about your character than your reputation. Character is what you
are, reputation merely what others think you are." - John Wooden
-Original Message-
From:
It was a while back, but I'm pretty sureI have used _BPX_SHAREAS and the spawn
syscall from REXX, to run Java in the same address space.
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf
Of Lionel B Dyck
Sent: Monday, November 4, 2019
Thats fine. However, java itself runs under unix...
ITschak
בתאריך יום ב׳, 4 בנוב׳ 2019, 14:30, מאת Steve Austin <
steve.aus...@macro4.com>:
> It was a while back, but I'm pretty sureI have used _BPX_SHAREAS and the
> spawn syscall from REXX, to run Java in the same address space.
>
> -Ori
Fair enough. My reply was in response to the comment (which I apparently forgot
to quote) about having 3x real memory for paging space.
Depending on where the problem is, even if you're not paging, more memory could
be beneficial if the system was configured to use it appropriately. I.E. if the
Enclaves were supposed to be an exception to this rule. First, SRB's in an
enclave are interruptible. Second, for dispatching purposes, SRB's and
TCB's placed in an enclave are excluded from the address space. The system
will select either an enclave or address space for processing. Time
slic
On Mon, 4 Nov 2019 05:20:14 -0600, Lionel B Dyck wrote:
>Are there any hints/tips to improve the performance of an ISPF application
>that uses BPXWUNIX? In running on my normal system the performance is very
>slow compared to a friends system running on a zPDT.
>
I've had bad performance due to
From days of yore,
IBM has generally recommended 2:1 virtual to real storage. I have pushed this
to 3:1 without major issues.
The paging subsystem should be configured accordingly page slots approx. 3x
available real.
If you are actually going to be doing serious paging, (> +/- 20 pps) the
l
On Mon, 4 Nov 2019 14:04:02 +, Allan Staller wrote:
>IBM has generally recommended 2:1 virtual to real storage. I have
>pushed this to 3:1 without major issues.
>The paging subsystem should be configured accordingly page slots
>approx. 3x available real.
>
>If you are actually going to be d
Bright #pragma runoptions wrapped up in strings
These are a few of my favorite things .
>
> LOL! Wrap them in a string!
>
> #pragma runopts("TEST(,,,TCPIP&192.168.1.164%8001:*)")
>
>> On 2019-11-04 12:37 PM, Joseph Reichman wrote:
>> #pragma runopts(TEST(,,,TCPIP&192.168.1.164%8001:*))
>
Have a CEETEST in my amode 64 C program configured z/os explorer I’ll find out
shortly if debug tool takes off in z/os explorer
> On Nov 4, 2019, at 9:57 AM, Chris Hoelscher wrote:
>
> Bright #pragma runoptions wrapped up in strings
> These are a few of my favorite things .
>
>>
>>
The number of local page slots provided need to support the virtual storage in
use, so according to the algorithms posted below. Yes.
If 30 TB of virtual is in use, about 30TB of page slots should be provided
*(you can fudge things here at your own risk)*.
Applying the 30% utilization rule to th
Many of the newer z boxes come with Flash memory now, in 1.5Tb increments. Not
sure if that is a separately chargeable feature or not.We configure the
appropriate amount that shows up as SCM memory when doing a D ASM, and a drill
down of D ASM,SCM. I'm though I read somewhere I saw that S
Gentle Readers,
We receive a IEC161I 009-0663 (COND CODE 0008) everytime we execute a
batch job which uses DCOLLECT: The error occurs while reading the MCDS.
I checked the IBM documentation and I got more confused. I found this tidbit
of info:
This is a known condition when running IDCAM
In the LISTCAT output look for "RLS IN USE", it should show YES or NO.
Thanks..
Paul Feller
AGT Mainframe Technical Support
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf
Of esmie moo
Sent: Monday, November 04, 2019 9:58 AM
To: IBM-MAI
Hello Group,
We are in the process of installing DB2 v12 and I find minimum memory limit
in DSN1DBM1 address is 16M . Can you please help me to understand if this
is minimum memory requiremen by this address space which is different then
db2 v11 , which was only 4M.
Can you please suggest.
-
It’s probably MEMLIMIT - which governs VIRTUAL Storage. And it will be in
GB.
If that’s what Db2 V12 Installation says, you’d better make sure it has it
- whether through JCL or exit or PARMLIB or whatever.
But Db2 will only take what it needs - with the proviso that you can tune
what it needs.
On Mon, 4 Nov 2019 at 07:32, ITschak Mugzach wrote:
On Mon, 4 Nov 2019 at 07:30, Steve Austin wrote:
> It was a while back, but I'm pretty sureI have used _BPX_SHAREAS and the
> spawn syscall from REXX, to run Java in the same address space.
Thats fine. However, java itself runs under unix..
On Sun, 3 Nov 2019 at 20:31, Timothy Sipples wrote:
> Matt Hogstrom wrote:
> >https://github.com/walmartlabs/zUID
> >Courtesy of Walmart
>
> Tony Harminc wrote:
> >Wouldn't want to bump into that pending patent from Walmart...
>
> Walmart licensed the code they're sharing under the Apache License
The Java program does not run as a main program, but need a shell
environment such as the one supplied by BPXBATCH. So actually the java
program already runs under a unix thread. this is not the case with Lionel
performance issue, as his code runs under TSO and requires an extra asid to
run the uni
On Mon, 4 Nov 2019 at 14:07, ITschak Mugzach wrote:
> The Java program does not run as a main program, but need a shell
> environment such as the one supplied by BPXBATCH. So actually the java
> program already runs under a unix thread. this is not the case with Lionel
> performance issue, as his
On Mon, 4 Nov 2019 14:18:30 -0500, Tony Harminc wrote:
>
>Many programs that are customarily started from a shell prompt don't
>have to be. Often this is just UNIXism at play. I have had success in
>the past copying UNIX executable files (i.e. Program Objects in UNIX
>files) into PDSE members, and
On Mon, 4 Nov 2019 05:20:14 -0600, Lionel B Dyck wrote:
>Are there any hints/tips to improve the performance of an ISPF application
>that uses BPXWUNIX? In running on my normal system the performance is very
>slow compared to a friends system running on a zPDT.
>
See also:
https://www.ibm.c
MVS has provided such a mechanism forever using
1. a program that issues a WTO
2. conditional execution on the WTO step with COND=EVEN or ONLY (depending on
the logic required)
Most shops already have a more or less sophisticated (security wise) program to
issue a WTO. Conditional JCL is free
I used to be on one of the IBM Invention Development Teams and the rules on
patents changed a few years ago. Prior to the change, the protection date
was date of invention. After the change, it became the patent filling date.
If you publish anything about the patent you intend to file, anyone can
t
On 11/4/2019 9:39 AM, Tom Marchant wrote:
On Mon, 4 Nov 2019 14:04:02 +, Allan Staller wrote:
IBM has generally recommended 2:1 virtual to real storage. I have
pushed this to 3:1 without major issues.
The paging subsystem should be configured accordingly page slots
approx. 3x available real
Jesse/Brian,
Same here in most shops someone had all written a WTO Pgm.
Scott
On Mon, Nov 4, 2019 at 4:02 PM Jesse 1 Robinson
wrote:
> MVS has provided such a mechanism forever using
>
> 1. a program that issues a WTO
>
> 2. conditional execution on the WTO step with COND=EVEN or ONLY (dependi
It would be nice if there were a way to get the condition codes into the WTO
program, something like (don't try this at home):
//WTO1 EXEC PGM=WTO,COND=EVEN,
// PARM='Job FOOBAR terminating. Highest CC was &MAXCC'
I seem to recall a discussion here about retrieving the last and maximum job
CC'
Step and job end status are very accessible. Look at SYS1.SAMPLIB(SMFEXITS).
Factors like step and maximum condition codes; step and job CPU totals; and I/O
counts can be extracted from control blocks and displayed in WTOs. The code
versions we use have been tailored over the years so that our m
I have ten years of commercial product experience with SMF exit development.
:-) I am super familiar with SMF 30 subtypes 4 and 5 (and all the others). I
have a yellow card.
I was referring to a program running as a jobstep retrieving condition codes
for its own job. Is there an MVS service or
The COND=EVEN/ONLY will not work if the job hits a *JCL error*.
On 04/11/2019 23:56, Charles Mills wrote:
> It would be nice if there were a way to get the condition codes into the WTO
> program, something like (don't try this at home):
>
> //WTO1 EXEC PGM=WTO,COND=EVEN,
> // PARM='Job FOOBAR
On 2019-11-05 8:52 AM, Charles Mills wrote:
I was referring to a program running as a jobstep retrieving condition codes
for its own job. Is there an MVS service or control block chain that readily
gives that information?
IIRC, the SCT (Step Control Table) has this information and the SCT
e
I got a private note that says JSCB to JCT to SCT to SCTX.
It was just kind of pseudo-intellectual musing on my part. I have no need.
Sounds like the problem of "WTO message at the end of [a job]" (the OP) via a
jobstep is eminently doable. The only problem unsolvable in that way is JCL
error.
This sure sounds like a discussion from the 1980s. Operators? Consoles?
Those are still a thing?
Try out the new // NOTIFY statement, and send an email to whoever needs to
wake up.
sas
On Mon, Nov 4, 2019 at 8:37 PM CM Poncelet wrote:
> The COND=EVEN/ONLY will not work if the job hits a *JCL
To net it out, thanks again to Walmart (and to the specific authors at
Walmart) for creating and sharing this code. Bravo.
Timothy Sipples
IT Architect Executive, Industry Solutions, IBM Z & Li
Of course there is a way, all you need to do is process the SCT from your
program. You will get all steps up to the one you are currently in only
because it has not ended yet. But if you are running as the last step it
wouldn't matter. That's not how I do it in SyzMPF/z, but it "could" be don
That's the reason that SyzMPF/z sells so well, it creates the email (or test
message) at the end of the job whether it got a JCL error Abend or normal end
(with any and all condition codes), in a nice little email in either HTML or
plain text (or sms text message) and can even send any or all pa
I use a REXX script which we use to record cond codes from test jobs.
main:
cvt = ptr(16)
tcbp = ptr(cvt)
tcb = ptr(tcbp + 4)
jscb = ptr(tcb + 180)
ssib = ptr(jscb + 316)
tct = ptr(tcb + 164)
lct = ptr(tct + 152)
jct = ptr(lct + 16)
jobid = stg(ssi
44 matches
Mail list logo
