Thanks Al for additional info. I'm embarrassed to admit that I don't know if 
DASD-only logging requires any CF at all. I thought not but hesitated to say 
that in my post. 

.
.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler 
SHARE MVS Program Co-Manager
626-302-7535 Office
323-715-0595 Mobile
[email protected]

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Nims,Alva John (Al)
Sent: Tuesday, January 27, 2015 10:39 AM
To: [email protected]
Subject: Re: Health Checker and SMF

As a possible source of information about migrating SMF collection to a 
DASD-Only log stream you could check IBM's "z/OS V1R10.0 System z Platform Test 
Report for z/OS and Linux Virtual Servers" report from June of 2009, "Chapter 
4. Migrating from SMF data set recording to log stream logging".  Found in 
zPET's archives:

http://www-03.ibm.com/systems/services/platformtest/servers/systemz_archives.html

I helped in the testing of the SMF log stream conversion and some of the REXX 
code In the chapter is from me and sorry, when I gave the code, I did not 
expect it to get stuck in the document like that so the statement: "PIPE STACK 
| PAD 80 | CHOP 80 | > DDNAME=SMFCNTL"
Just convert it to an EXECIO to write to DDNAME=SMFCNTL, I was using the "PIPE" 
command to do a lot of things in those days.


Al Nims
Systems Admin/Programmer 3
Information Technology
University of Florida
(352) 273-1298

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of J O Skip Robinson
Sent: Tuesday, January 27, 2015 11:49 AM
To: [email protected]
Subject: Re: Health Checker and SMF

Before imploring IBM to add new function to an already deprecated SMF 
collection mechanism, I suggest giving SMF logger a try. Even with limited 
coupling facility resources, you can set up DASD-only log streams--although I 
have no experience with these. SMF logger is not bullet proof, but it's more 
resilient in episodes of high usage. And this situation no longer occurs:

All MANx datasets have somehow been marked closed while only partially full, 
sometimes only a few percent. At IPL time, SMF warns that no data sets are 
available and begins buffering. That warning can easily get lost in the 
cacophony of messages issued during IPL. Depending on your tape technology, you 
may not be able to offload even a single MANx before you start losing data. SMF 
logger, which can get flooded as well, is much less susceptible to a 'full' 
condition and loss of data. 

.
.
.
J.O.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
626-302-7535 Office
323-715-0595 Mobile
[email protected]

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Vernooij, CP (ITOPT1) - KLM
Sent: Tuesday, January 27, 2015 7:28 AM
To: [email protected]
Subject: Re: Health Checker and SMF

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Robert A. Rosenberg
Sent: 27 January, 2015 16:04
To: [email protected]
Subject: Re: Health Checker and SMF

At 14:46 +0000 on 01/27/2015, Vernooij, CP (ITOPT1) - KLM wrote about
Re: Health Checker and SMF:

>And to your question about HC for SMF: I pleaded earlier and repeat it
>now: we should restrict HC to report only on unhealthy situations and 
>try to avoid that HC floods us with all kinds of thoughts, 
>recommendations, considerations etc. that could fill an entire Redbook.
>So for SMF: have at least 3 man datasets is a check that could avoid 
>bringing down SMF.
>
>Kees.

Another Check that could be reported is if the last SMF Dataset is the active 
one (ie: All the others needs to be dumped since there is no empty dataset to 
be rolled over to).

----------------------------------------------------------------------

SMF is already telling you when this is the case and you could/should have 
automated this message.
In my opinion, this not an unhealthy system configuration situation for HC to 
report. Besides, catching these events requires almost a realtime monitor, 
which HC is not.

Kees.  

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to