Re: Incomplete monitor data: SAMPLE CONFIG size too small
I have yet to see the 5.3 version of the manual, but that section in the 5.2 version is patently incorrect. When I tried to open an apar to get the right numbers (I didn't go the RCF route because getting it corrected in a future edition wasn't going to help with the problem I was having) I was counselled that that wouldn't help, but an internal issue would be opened and I could expect that relief would be forthcoming. That was several months ago and I'm still waiting... (I eventually arrived at happy/reasonable allocations for the monitor dcss through trial and retrial, as others have done; but that doesn't help the next one to run into the same wall.) Mike Harding EDS VM National Capability 134 El Portal Place Clayton, Ca. USA 94517-1742 * phone: +01-925-672-4403 * Fax: +01-925-672-4403 * mailto:[EMAIL PROTECTED] * mailto:[EMAIL PROTECTED] (personal) Note: For 2007, I am off on Fridays with even Julian dates and Mondays with odd ones. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Alan Ackerman Sent: Tuesday, November 27, 2007 9:02 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Incomplete monitor data: SAMPLE CONFIG size too small See Calculating the Space Needed for the Saved Segment in z/VM V5R3.0 Performance. Alan Ackerman Alan (dot) Ackerman (at) Bank of America (dot) com
Re: Incomplete monitor data: SAMPLE CONFIG size too small
For those with esamon, esamon creates the dcss and starts the monitor correctly to avoid such problems. Harding, Mike wrote: I have yet to see the 5.3 version of the manual, but that section in the 5.2 version is patently incorrect. When I tried to open an apar to get the right numbers (I didn't go the RCF route because getting it corrected in a future edition wasn't going to help with the problem I was having) I was counselled that that wouldn't help, but an internal issue would be opened and I could expect that relief would be forthcoming. That was several months ago and I'm still waiting... (I eventually arrived at happy/reasonable allocations for the monitor dcss through trial and retrial, as others have done; but that doesn't help the next one to run into the same wall.) Mike Harding EDS VM National Capability 134 El Portal Place Clayton, Ca. USA 94517-1742 * phone: +01-925-672-4403 * Fax: +01-925-672-4403 * mailto:[EMAIL PROTECTED] * mailto:[EMAIL PROTECTED] (personal) Note: For 2007, I am off on Fridays with even Julian dates and Mondays with odd ones. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Alan Ackerman Sent: Tuesday, November 27, 2007 9:02 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Incomplete monitor data: SAMPLE CONFIG size too small See Calculating the Space Needed for the Saved Segment in z/VM V5R3.0 Performance. Alan Ackerman Alan (dot) Ackerman (at) Bank of America (dot) com
Re: Incomplete monitor data: SAMPLE CONFIG size too small
And believe me, I'd much rather we were using esamon. Mike Harding EDS VM National Capability 134 El Portal Place Clayton, Ca. USA 94517-1742 * phone: +01-925-672-4403 * Fax: +01-925-672-4403 * mailto:[EMAIL PROTECTED] * mailto:[EMAIL PROTECTED] (personal) Note: For 2007, I am off on Fridays with even Julian dates and Mondays with odd ones. -Original Message- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of barton Sent: Tuesday, December 04, 2007 1:50 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: Incomplete monitor data: SAMPLE CONFIG size too small For those with esamon, esamon creates the dcss and starts the monitor correctly to avoid such problems.
Re: Incomplete monitor data: SAMPLE CONFIG size too small
On Mon, 26 Nov 2007 14:19:07 -0600, RPN01 [EMAIL PROTECTED] wrote: This matches what I got to by trial and error. I¹ve ended up with the mondcss segment from 9000-9fff, and a sample config size of 1500. It see ms to be working without error now. This started with installing Omegamon X E for z/VM and Linux, and I still need to see if that is receiving any dat a at this point. -- .~.Robert P. Nix Mayo Foundation /V\RO-OE-5-55 200 First Street SW / ( ) \ 507-284-0844 Rochester, MN 55905 ^^-^^ - In theory, theory and practice are the same, but ³Join the story. .. Ride Ural.² in practice, theory and practice are different. See Calculating the Space Needed for the Saved Segment in z/VM V5R3.0 Pe rformance. Or consult the documentation for your performance monitor. There is no penal ty for making the Sample portion of the DCSS too big -- CP always starts over from the bott om on each sample interval. There may be a paging cost if you make the EVENT area too big, since CP keeps adding event data to the end until it is full before wrapping. Alan Ackerman Alan (dot) Ackerman (at) Bank of America (dot) com
Re: Incomplete monitor data: SAMPLE CONFIG size too small
What I've had to do in the past is increase the DCSS size and the SAMPLE CONFIG size. We run with a SAMPLE CONFIG size of 1250. The MONDCSS is set to BEGPAG of 09000 and a ENDPAG of 09FFF. We had to go to 1250 size when we added a large amount of new DASD to the VM lpars. Paul From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of RPN01 Sent: Monday, November 26, 2007 10:57 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Incomplete monitor data: SAMPLE CONFIG size too small I realize that I've been here before, and I remember that the solution had nothing to do with the monitor command or anything else obvious... But I've slept since then and have come up short of the solution. When I start the Performance Toolkit service machine, it gets this exception: FCXPMN446E Incomplete monitor data: SAMPLE CONFIG size too small What is the proper way to track down a solution to this message? I increased the sample config size from 241 to 700, and still get the same error. Increasing it to 1000 gives two errors: FCXPMN446E Incomplete monitor data: SAMPLE CONFIG size too small FCXPMN446E Incomplete monitor data: DCSS size too small I remember that last time, the solution had nothing to do with the DCSS at all, but I've lost any notes I might have had on it. Plus, this time the problem started with the SAMPLE CONFIG message instead of the DCSS message, so I'd like to get some input before I make any major changes... -- .~.Robert P. Nix Mayo Foundation /V\RO-OE-5-55 200 First Street SW / ( ) \ 507-284-0844 Rochester, MN 55905 ^^-^^ - In theory, theory and practice are the same, but Join the story... Ride Ural. in practice, theory and practice are different.
Re: Incomplete monitor data: SAMPLE CONFIG size too small
On Nov 26, 2007 6:54 PM, Feller, Paul [EMAIL PROTECTED] wrote: What I've had to do in the past is increase the DCSS size and the SAMPLE CONFIG size. We run with a SAMPLE CONFIG size of 1250. The MONDCSS is set to BEGPAG of 09000 and a ENDPAG of 09FFF. We had to go to 1250 size when we added a large amount of new DASD to the VM lpars. Indeed. The complaint is about the sample configuration size that needs to be larger, but often you cannot really do so without making the MONDCSS larger. The sample provided by z/VM is too small for most installations. There is no penalty for making it a bit too large, and we find that 64M works for most folks. The size of the sample configuration area in the MONDCSS is set when the monitor is started (the default of 241 pages is too small for many). My rough estimate is one page for every 10 devices defined. Your performance monitor will tell you how much is really being used. Rob -- Rob van der Heij Velocity Software, Inc http://velocitysoftware.com/
Re: Incomplete monitor data: SAMPLE CONFIG size too small
This matches what I got to by trial and error. I¹ve ended up with the mondcss segment from 9000-9fff, and a sample config size of 1500. It seems to be working without error now. This started with installing Omegamon XE for z/VM and Linux, and I still need to see if that is receiving any data at this point. -- .~.Robert P. Nix Mayo Foundation /V\RO-OE-5-55 200 First Street SW / ( ) \ 507-284-0844 Rochester, MN 55905 ^^-^^ - In theory, theory and practice are the same, but ³Join the story... Ride Ural.² in practice, theory and practice are different. On 11/26/07 11:54 AM, Feller, Paul [EMAIL PROTECTED] wrote: What I've had to do in the past is increase the DCSS size and the SAMPLE CONFIG size. We run with a SAMPLE CONFIG size of 1250. The MONDCSS is set to BEGPAG of 09000 and a ENDPAG of 09FFF. We had to go to 1250 size when we added a large amount of new DASD to the VM lpars. Paul