Hi Victor,
Best answer is get to z/OS R6 where the SMF buffering just works better and is
easily expanded. Since you said you were currently on R4 though you can do
this with a USERMOD. As Walt suggested get yourself some more buffers. Paul
Gillis kindly shared his on the MXG-L list a while ago.
You need to carefully review OW56001 and make sure you have enough resources to
support whatever you specify!
We set our SMF buffer sizes on OS/390 2.10 and z/OS 1.4 as follows, and have
not since seen any SMF buffer problems since applying the USERMOD.
++ USERMOD(LIEE001) REWORK(2004163)
DESC(Zaps to SMFLIM to reset SMF Default Values).
++VER(Z038) FMID(HBB7707) PRE(UW94068).
++ZAP(IEEMB822).
NAME SMFLIMS
VER 0000 00000008 * Initial buffer size (8 meg)
VER 0004 00000008 * Buffer increment size (8 meg)
VER 0008 00000080 * Maximum buffer size (128 meg)
VER 000C 00000019 * Buffer warning level (25%)
REP 0000 00000080 * Initial buffer size (128 meg)
REP 0004 00000020 * New Increment size = 32 meg
REP 0008 00000400 * New Maximum size = 1024 meg
REP 000C 00000032 * New Warning level = 50%
Best Regards,
Sam Knutson, GEICO
Performance and Availability Management
mailto:[EMAIL PROTECTED]
(office) 301.986.3574
Murphy's Computer Law 39: An ounce of image is worth a pound of performance.
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Walt
Farrell
Sent: Wednesday, January 18, 2006 10:26 AM
To: [email protected]
Subject: Re: How SMFDUMP works?
On 1/17/2006 2:38 AM, Víctor de la Fuente wrote:
> Our problem is also SMF starts losing data before all files are full:
>
> IEE979W SMF DATA LOST - NO BUFFER SPACE AVAILABLE TIME=08:19:30 ...
> D SMF
> IEE974I 08.20.04 SMF DATA SETS 177
> NAME VOLSER SIZE(BLKS) %FULL STATUS
> P-SYS1.MAN1E1 CGEHAA 43200 100 CLOSE PENDING
> S-SYS1.MAN1E2 CGEHAB 43200 16 ACTIVE
> S-SYS1.MAN1E3 CGEHAC 43200 0 ALTERNATE
> S-SYS1.MAN1E4 CGEHAD 43200 0 ALTERNATE
> S-SYS1.MAN1E5 CGEHAE 43200 0 ALTERNATE
>
> So sys1.man1e1 does not reach "Dump Required" state, and also SMF has
> used all of his space available...
The fact that SYS1.MAN1E1 is "CLOSE PENDING" rather than "DUMP REQUIRED"
does not seem relevant to your problem. You DO have an ACTIVE data set:
SYS1.MAN1E2, which should be able to record the records. Thus, the switch has
occurred, even if the old data set is not yet available for dumping.
Assuming that SYS1.MAN1E1 does go to "DUMP REQUIRED" and you can dump it,
before you fill SYS1.MAN1E5, then your only problem is that your in-storage
buffers are filling before they can be written to SYS1.MAN1E2, and for that you
need more buffers or faster DASD, I think.
Walt
====================
This email/fax message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution of this
email/fax is prohibited. If you are not the intended recipient, please
destroy all paper and electronic copies of the original 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
