In the absence of CF Sizer assistance ;-( I looked at the samples from IBM
and thought they seemed awfully big. A structures occupies dedicated real
storage, after all; I tend to be stingy with it. After consulting with our
SMF data caretaker who saw no obvious advantage is splitting records at the
collection point, I set up only a single CF structure like this that
captures all records:
STRUCTURE
NAME(IFASMF_DEFAULT) /* SMF structure for unspecified types */
INITSIZE(10000)
SIZE(50000)
Despite occasional IXC585E 'structure full' messages, I haven't increased
it from the INITSIZE value. This structure supports only one system.
Another member is 1.7 (no can do), and the third is also 1.9 but a
'bronze-plex' member that does not share DASD. Haven't quite figured out
how handle that puppy yet.
An interesting quirk I just noticed: most (for today, all!) IXC585E
messages are being issued by the other 1.9 member, which still uses MANx
recording. No messages on the 1.7 guy or on the guy actually using System
Logger.
We have several non-parallel-sysplex systems that are candidates for
DASD-only logging but haven't ventured down that road as yet. Book says it
works.
Jim Holloway
<[EMAIL PROTECTED]
E.COM> To
Sent by: IBM IBM-MAIN@BAMA.UA.EDU
Mainframe cc
Discussion List
<[EMAIL PROTECTED] Subject
.EDU> Re: SMF in System Logger
04/14/2008 11:20
AM
Please respond to
IBM Mainframe
Discussion List
<[EMAIL PROTECTED]
.EDU>
On 04/14/2008 01:20 PM, Skip Robinson wrote:
> I presented a user session at SHARE Orlando based on experience with our
> z/OS 1.9 sandbox image. In the meantime, we have migrated SMF Logger to
our
> development system, which gets very busy and therefore creates way more
SMF
> data than the sandbox does. Everything is working fine. I still miss the
<Snip>
Skip,
Thanks for the feedback. I couldn't find a manual reference for
SMF Structure names either. But I am curious as to the structure sizing
you're using
for your development environment(If you're using the CF). I have three
structures defined, type 30s, type 80s,
and everything else. The lab environment I'm on is our smallest lab
sysplex used for DASD Mirroring
testing. So there are no type 110s or type 101s and in reality not much
happens in this plex.
Our primary labs have CICS and DB/2 so the volume should go up. In
production, which is fairly busy,
our capacity folks have reserved 8Gig in our external CFs for SMF. The
SYSPLEX is a 5 system sysplex and we're
looking at having five structures/logstreams per system. Also, if you are
using the CF, are you duplexing the data
into staging datasets? We are considering this as a "Belt and Suspenders"
solution to help insure against dataloss.
Jim Holloway
Sr. Software Systems Engineer
Operating Systems & Automation - RISC - Troy, NY
(518)285-2646 [Voice]
(518)281-9686 [Cell Phone]
(419)730-0857 [FAX]
MetNet Lotus Notes: mailto:jholloway
Internet: mailto:[EMAIL PROTECTED]
The information contained in this message may be CONFIDENTIAL and is for
the intended addressee only. Any unauthorized use, dissemination of the
information, or copying of this message is prohibited. If you are not the
intended addressee, please notify the sender immediately and delete this
message.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
