1 WWW.IDUGDB2-L.ORG /home/listserv/home/db2-l June 2009, week 3
2 53 24_[MVS] Quiesce Sequencing11_Nick Cianci19_ncianci@AU1.IBM.COM31_Mon, 15 Jun 2009 14:34:31 +1000619_ISO-8859-1 Hi Ladies & Gents, we've had some intermittent difficulties with Quiesces lately. Some of our issues might be resolved if we could determine a way to get quiesce to take the more volatile locks first, there by reducing the likelihood of an application-utility deadlock scenario.
A few years ago ( under v7, v8 now) we were experimenting with the order in the LISTDEF. However, whilst the messages were spat out in the correct sequence, anecdotal evidence at the time seemed to imply that the drain locks were acquired in an entirely different sequence approximating a LIFO sequence (push and pop [...]
56 204 47_Redbook DB2 9 for z/OS Distributed Architecture10_DB2usa !!!19_db2usa3@HOTMAIL.COM31_Mon, 15 Jun 2009 04:38:17 -0400767_Windows-1252
Hi DB2 user,
Here is a non-profit Blog about DB2 for z/OS (IBM mainframes): http://db2usa.blogspot.com
Last update on Monday, June 15th 2009
IBM just published a draft version of a new Redbook, "DB2 9 for z/OS: Distributed Architecture, SG24-6952-01".
Abstract Distributed Relational Database Architecture™ (DRDA® ) is a set of protocols that permits multiple local and remote database systems and application programs, to work together. Any combination of relational database management products that use DRDA can be connected to form a distributed relational database management system. DRDA coordinates communication between systems by defining what can be exchanged and how it [...]
261 46 49_why occures DSNR035I in batch only using select??34_=?KS_C_5601-1987?Q?Donghyun_Jung?=19_jdonghyun@GMAIL.COM31_Mon, 15 Jun 2009 09:27:17 +0000384_ks_c_5601-1987 hi, all~
i have a long running batch that only issues "SELECT ... WITH UR" queries using cursor...
But, i got some warning messages(DSNR035I) about this batch in DB2 v8 master log on z/OS. so. i don't understand why DSNR035I was occured..
DSNR035I indicates that in checkpoint, DB2 is aware of an uncommited ur which has an inflight status... [...]
308 101 61_AW: [DB2-L] why occures DSNR035I in batch only using select??12_Peter, Georg15_G.Peter@DZBW.DE31_Mon, 15 Jun 2009 11:47:26 +0200687_iso-8859-1 PQ94994: DSNR035I / MSGDSNR035I IS INCORRECTLY ISSUED:
See http://www-01.ibm.com/support/docview.wss?uid=swg1PQ94994
HTH.
With kind regards - mit freundlichen Gruessen, G e o r g H . P e t e r ------------------------------------------------------------------- Datenzentrale Baden-Wuerttemberg Development and Product Support (E3) Krailenshaldenstrasse 44, 70469 Stuttgart, Germany, Europe e:mail G.Peter@dzbw.de Phone 0049-711-8108-271 PC-Fax 004971189696071 Internet (only in german language):http://www.dzbw.de ---------------------------------------------------------------------- THINK GREEN - DON'T PRINT THIS EMAIL UNLESS YOU REALLY NEED TO ! [...]
410 195 65_Re: AW: [DB2-L] why occures DSNR035I in batch only using select??34_=?KS_C_5601-1987?Q?Donghyun_Jung?=19_jdonghyun@GMAIL.COM31_Mon, 15 Jun 2009 10:26:02 +0000402_ks_c_5601-1987 hi, georg....
thanks for fast reply...
but this ptf already applied in our system....
our system status is as follows;
CSI CROSS-ZONE QUERY - ENTRY SELECTION Row 1 of 3 ===> SCROLL ===> PAGE
Entry Type: SYSMOD Entry Name: UK00898
To return to the previous panel, enter END .
To select an entry from a zone, enter S next to the zone. [...]
606 88 28_Re: [MVS] Quiesce Sequencing13_Bell, Raymond20_Raymond_Bell@BMC.COM31_Mon, 15 Jun 2009 14:51:15 +0200554_iso-8859-1 Hi Nick,
Not sure I qualify for either of your two addressed groups or not (well, I'm pretty sure I don't qualify for the first one) but I'll chip in anyway.
I don't know in what order Quiesce decides to try to Quiesce objects in a Listdef but, to bend a reference, the fastest Quiesce is no Quiesce. Do you really need to take the Quiesce i.e. run the Quiesce utility? I mean, as opposed to merely identifying a quiet point. 'Cause if it's the latter there are products out there that will do that for you, after the fact. [...]
695 23 30_Re: REXX, SQL and FOR BIT DATA15_Leland Goldston23_lgoldston@EARTHLINK.NET31_Mon, 15 Jun 2009 14:01:41 +0000638_windows-1252 We are DB2 V8 in NFM. I am not seeing any symptoms what so ever other than +100 SQLCODE upon first fetch. I have taken the LIKE out of the select/cursor and code works beautifully.
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG North America Attendee Testimonial- "The round table SIGs I attended were great. The IBM presenters were very helpful and professional." _____________________________________________________________________
719 42 30_Re: REXX, SQL and FOR BIT DATA9_Don Leahy19_don.leahy@LEACOM.CA31_Mon, 15 Jun 2009 10:29:57 -0400675_ISO-8859-1 Could this be a UNICODE vs. EBCDIC issue? FOR BIT DATA normally implies no translation.
I have no advice to offer because we are still running V8 in CM. :-(
On Mon, Jun 15, 2009 at 10:01 AM, Leland Goldston wrote: > We are DB2 V8 in NFM. I am not seeing any symptoms what so ever other > than +100 SQLCODE upon first fetch. I have taken the LIKE out of the > select/cursor and code works beautifully. > > _____________________________________________________________________ > > * IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU * > > _____________________________________________________________________ > [...]
762 49 30_Re: REXX, SQL and FOR BIT DATA9_Chris Tee23_Chris.Tee@SENTENIAL.COM31_Mon, 15 Jun 2009 15:44:48 +0100538_us-ascii Leland
In NFM the SQL in SYSPACKSTMT is in Unicode so it won't match your ebcdic LIKE clause. I didn't see your original post but you have something like SELECT CAST(CAST(STMT AS VARCHAR(3500) CCSID 1208) AS VARCHAR(3500) CCSID EBCDIC) FROM SYSIBM.SYSPACKSTMT in your query?
Chris
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Leland Goldston Sent: 15 June 2009 15:17 To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] REXX, SQL and FOR BIT DATA [...]
812 24 36_SV: [DB2-L] [MVS] Quiesce Sequencing13_Hanne Lyssand20_Hanne.Lyssand@VPS.NO31_Mon, 15 Jun 2009 16:57:40 +0200560_us-ascii In DB2 V9 DB2 will resolve all in flight UR's so the need for QUIESCE is even less if you V9.
Hanne
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG North America Attendee Testimonial- "The round table SIGs I attended were great. The IBM presenters were very helpful and professional." _____________________________________________________________________
837 78 30_Re: REXX, SQL and FOR BIT DATA9_Mike Bell21_mbell11a1@VERIZON.NET31_Mon, 15 Jun 2009 10:03:19 -0500544_us-ascii UM, not true - the SQL text in SYSPACKSTMT is in whatever character encoding the origional DBRM was in. If the DBRM is precompiled with NEWFUN(YES) then the SQL text is in unicode. What this means is that a single SQL statement can NOT retrieve information from SYSPACKSTMT without a substantial amount of work. The other part of the problem is that there is other data in SYSPACKSTMT that truly is binary data, not character and a blanket conversion as if everything was unicode will give error messages for every row processed. [...]
916 477 65_Re: AW: [DB2-L] why occures DSNR035I in batch only using select??15_Chris Hoelscher21_choelscher@HUMANA.COM31_Mon, 15 Jun 2009 12:50:44 -0400612_ISO-2022-KR $)Cyhe fix may have been applied to your install SDSNLOAD - has it been propagated into your runtime SDSNLOAD?
Chris Hoelscher Senior IDMS & DB2 Database Administrator Humana Inc 502-476-2538 choelscher@humana.com
you only need to test the programs that you want to work correctly
From: Donghyun Jung To: DB2-L@WWW.IDUGDB2-L.ORG Date: 06/15/2009 06:27 Subject: Re: [DB2-L] AW: [DB2-L] why occures DSNR035I in batch only using select?? Sent by: DB2 Data Base Discussion List [...]
1394 1835 65_Re: AW: [DB2-L] why occures DSNR035I in batch only using select??16_Broyles, Carol L27_Carol.L.Broyles@ACS-INC.COM31_Mon, 15 Jun 2009 11:57:23 -0500534_iso-2022-jp Like Chris, I never trust what the SMP zone is telling me. Run a MEPL to verify that it $B!G (Js on.
Carol L. Broyles
Infrastructure Mgt. Consultant
Commercial Solutions
Office Phone: 937-495-4003
carol.l.broyles@acs-inc.com
Affiliated Computer Services Inc.
CONFIDENTIALITY NOTICE: This e-mail message, including any attachments,
is for the sole use of the intended recipient(s) and may contain confidential and [...]
3230 110 63_Re: Re-link to convert Stored Proc from SPAS to WLM not working4_Amar18_amar_moh@YAHOO.COM31_Mon, 15 Jun 2009 19:01:46 +0000343_windows-1252 Mike,
It was precompiled with ATTACH(CAF) but I had wrongly given INCLUDE DB2 (DSNELI) in SYSLIN DD in linked step. I think thats why DSNELI was showing up. Thanks for catching that.
So I did the precompile with ATTACH(CAF) again with following in Linked syslin DD after the object code. INCLUDE DB2LIB(DSNALI) [...]
3341 94 28_Re: [MVS] Quiesce Sequencing16_mocion@libero.it16_mocion@LIBERO.IT31_Mon, 15 Jun 2009 16:38:45 -0500572_ISO-8859-1 In V7 I read there was listdef optimization processing objects by size from largest to smallest (Bruni, 2003) to create a balanced distribution of objects across tapes/disks but I don't know if it's valid for quiesce.
Mx Scarpa
Bell, Raymond ha scritto: > Hi Nick, > > Not sure I qualify for either of your two addressed groups or not (well, I'm pretty sure I don't qualify for the first one) but I'll chip in anyway. > > I don't know in what order Quiesce decides to try to Quiesce objects in a Listdef but, to bend a reference, the [...]
3436 121 63_Re: Re-link to convert Stored Proc from SPAS to WLM not working9_Don Leahy19_don.leahy@LEACOM.CA31_Mon, 15 Jun 2009 18:09:56 -0400418_ISO-8859-1 As I mentioned earlier, ATTACH(CAF) causes CALL 'DSNHLI2' to be generated in your program source instead of the more common CALL 'DSNHLI'.
IMO, the best solution is to precompile the program without using ATTACH(CAF) and then relink using DSNRLI. Not using ATTACH(CAF) will cause the precompiler to generate calls to DSNHLI rather than DSNHLI2. Then, at Link edit time you just have to specify: [...]
3558 377 50_Concurrency during CREATE INDEX operations on z/OS22_Frank C. Fillmore, Jr.19_tfgpublic@GMAIL.COM31_Tue, 16 Jun 2009 07:29:46 -0400388_US-ASCII Okay Listers,
I sheepishly have to ask a rather basic question. A customer needs to implement a change in production and there's no convenient way to test it.
1. I need to create a new index on a table. If I issue a basic CREATE INDEX, will I still be able to issue INSERT/UPDATE/DELETE DML against that table while the index is being created? [...]
3936 460 62_AW: [DB2-L] Concurrency during CREATE INDEX operations on z/OS0_26_Walter.Janissen@ITERGO.COM31_Tue, 16 Jun 2009 13:53:01 +0200560_iso-8859-1 Hi
In the first case I would say, you are NOT able to issue inserts/updates/deletes, because I should the create handle an entry, which will be inserted in front of rows, the create-process has already read? So I think, there is at least an U-lock on the table-level.
In the second case, if the REBIULD-index is not running and you are on version 8 and the index is not unique and the SQL is dynamic, you can issue inserts/updates/deletes. If the rebuild is running, and you are not in version 9 using SHRLEVEL CHANGE, you can't. [...]
4397 269 27_ADMIN reminder PLEASE REAAD16_Galambos, Robert29_Robert.Galambos@COMPUWARE.COM31_Tue, 16 Jun 2009 07:57:29 -0400498_iso-8859-1 Hello DB2-L Subscribers,
This posting is something we (the list owners) are sending out about every couple of months. The purpose is to highlight information about DB2-L and also remind subscribers about certain information. Some information will be repeated to benefit new subscribers. We hope you will find the information useful, and save it for future reference. If you have feedback or suggestions, please contact us (the list owners) at DB2-L-Request@www.idugdb2-l.org [...]
4667 362 66_Re: AW: [DB2-L] Concurrency during CREATE INDEX operations on z/OS19_Jardine, Lawrence J19_JardineLJ@AETNA.COM31_Tue, 16 Jun 2009 09:29:59 -0400622_iso-8859-1 If you create the new index defer yes, you can run a shrlevel change reorg of the tablespace which will "rebuild" the index while allowing inserts/updates/deletes (except during last logapply and switch phase). I have used this technique many times.
Larry Jardine
Production DBA
________________________________
From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Walter.Janissen@ITERGO.COM Sent: Tuesday, June 16, 2009 7:53 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] AW: [DB2-L] Concurrency during CREATE INDEX operations on z/OS [...]
5030 924 39_[z/OS] RAND() function strange behavior14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Tue, 16 Jun 2009 14:08:05 -0400566_us-ascii Hello, all DB2 math-and-probability nerds! (The rest of you will be bored to tears by this, sorry, no need to read any further)
I'm looking for a pointer to a document or forum discussing the RAND() function, or someone who's done some experimenting with it. I have a problem that I've accidentally run across in using it for generating a binomial distribution where n=5 and p=.5 (five coin tosses) for a small (32000 or less) number of trials, and the resulting distribution is significantly different from the expected values. [...]
5955 251 69_Free IDUG Sponsored Webcast - DB2 9.7: Get the Inside Scoop - June 2419_McBride, Michael C.31_michael.mcbride@DOT.STATE.FL.US31_Tue, 16 Jun 2009 14:15:24 -0400513_us-ascii IDUG is proud to host a free webcast on IBM's new DB2 9.7 release. The one-hour webcast, DB2 9.7: Get the Inside Scoop, will be held Wednesday, June 24 at 12:00pm Eastern. Click to register for this free educational opportunity!
The webcast will feature IBM DB2 Evangelist Paul Zikopoulos, who will discuss the new features of DB2 9.7. Attendees will then be able to participate in a live Q&A session. [...]
6207 1099 43_Re: [z/OS] RAND() function strange behavior15_Blaicher, Chris22_Chris_Blaicher@BMC.COM31_Tue, 16 Jun 2009 13:56:51 -0500515_us-ascii From our V9 system
-------+---------+-------- BIN_SUM FREQUENCY -------+---------+-------- 0 1971 1 1879 2 11960 3 12138 4 2061 5 1991
It looks like random is not very random.
Chris Blaicher Phone: 512-340-6154 Mobile: 512-627-3803 ________________________________ From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Sevetson, Phil Sent: Tuesday, June 16, 2009 1:08 PM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] [z/OS] RAND() function strange behavior [...]
7307 502 43_Re: [z/OS] RAND() function strange behavior13_Michael Ebert18_mebert@AMADEUS.COM31_Tue, 16 Jun 2009 18:59:04 +0000484_ISO-8859-1 I'm pretty sure the problem is with your query, but I'm not at home with the WITH...AS... syntax. As a first test, create a table TEST(N1 INT, N2 INT, N3 INT, N4 INT, N5 INT) and do INSERT INTO TEST VALUES(ROUND(RAND(),0),...) 32000 times, then analyse that. As a first check, see whether you get a 50:50 distribution between 0 and 1. I don't have a DB2 that can do RAND() or WITH..., so you'll have to provide the input. I'll do some more thinking in the meantime :) [...]
7810 24 43_Re: [z/OS] RAND() function strange behavior11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Tue, 16 Jun 2009 19:08:08 +0000359_Windows-1252 >From our V9 system
>-------+---------+-------- > BIN_SUM FREQUENCY >-------+---------+-------- > 0 1971 > 1 1879 > 2 11960 > 3 12138 > 4 2061 > 5 1991
>It looks like random is not very random.
It's hard to say. The curve is still there. But, it's higher in the middle than it should be, and lowere at the ends. [...]
7835 1010 43_Re: [z/OS] RAND() function strange behavior13_Hampton, Kirk26_kirk.hampton@CAPGEMINI.COM31_Tue, 16 Jun 2009 15:10:16 -0400422_us-ascii Hello Phil, on my first try executing your SQL, on v8 NFM, I actually got worse than yours: ---------+---------+---------+ BIN_SUM FREQUENCY ---------+---------+---------+ 0 1281 1 2066 2 12778 3 11526 4 2990 5 1359
but on the fourth try, I got pretty close to your expectations: ---------+---------+---------+ BIN_SUM FREQUENCY ---------+---------+---------+ 0 978 1 5011 2 9929 3 10106 4 4984 5 992 [...]
8846 451 43_Re: [z/OS] RAND() function strange behavior14_Peter Backlund21_BacklundDB2@TELIA.COM31_Tue, 16 Jun 2009 21:49:18 +0200
9298 22 27_TRACKMOD on DB2 8/9 for zos17_Darren Kilpatrick31_darren.kilpatrick@WELLPOINT.COM31_Tue, 16 Jun 2009 19:59:35 +0000587_windows-1252 Just wondering if anyone has any info or stats on overhead of the Trackmod parameter (Yes vs. No) in a NON-data sharing envrionment.
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG North America Attendee Testimonial- "Each session I took away something that I could apply at my shop. I always learn something new." _____________________________________________________________________
9321 30 48_OSC problem - have to check tables after connect10_Rich Janni23_rich.janni@WAKEFERN.COM31_Tue, 16 Jun 2009 20:15:18 +0000422_windows-1252 I have OSC set up in my V8 subsystems. I'm using aliases to reference the explain tables via a secondary auth id. Everything works fine, except most times, after I connect to a subsystem, I only get 3 green lights. Explain enabled stays red, until I manually select 'use existing explain tables', then click 'CHECK TABLES'. This annoyance does not happen when I connect using sysadm or sysdba authority. [...]
9352 251 43_Re: [z/OS] RAND() function strange behavior9_Mike Bell21_mbell11a1@VERIZON.NET31_Tue, 16 Jun 2009 14:39:09 -0500569_US-ASCII I am making assumptions here - 1. the DB2 RAND function is the same LE CEERAN0 function for COBOL/PL1/C etc - it could be the RANDOM from cobol which is slightly different code. 2. the RAND function does not carry the seed number forward except on atomic SQL statements.
I have a performance / locking problem where the application was using RANDOM to get a random number for an id. The batch worked perfectly but the online was getting all kinds of collisions and locks. Turned out that the default seed for the RANDOM function was time of day to [...]
9604 322 43_Re: [z/OS] RAND() function strange behavior14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Tue, 16 Jun 2009 16:46:05 -0400579_us-ascii Thanks, Mike. That's very useful info.
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Mike Bell Sent: Tuesday, June 16, 2009 3:39 PM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] [z/OS] RAND() function strange behavior
I am making assumptions here - 1. the DB2 RAND function is the same LE CEERAN0 function for COBOL/PL1/C etc - it could be the RANDOM from cobol which is slightly different code. 2. the RAND function does not carry the seed number forward except on atomic SQL statements. [...]
9927 1345 43_Re: [z/OS] RAND() function strange behavior12_Hepp Shery C21_Shery.Hepp@SRPNET.COM31_Tue, 16 Jun 2009 14:56:06 -0700522_us-ascii Looks like it's IBM wide- here's the results on my V8 system
BIN_SUM FREQUENCY
----------- -----------
0 1526
1 3205
2 11099
3 11249
4 3317
5 1604
6 record(s) selected.
From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Blaicher, Chris Sent: Tuesday, June 16, 2009 11:57 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] [z/OS] RAND() function strange behavior [...]
11273 40 43_Re: [z/OS] RAND() function strange behavior11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Tue, 16 Jun 2009 22:27:03 +0000471_Windows-1252 >Looks like it’s IBM wide- here’s the results on my V8 system
>BIN_SUM FREQUENCY >----------- ----------- > 0 1526 > 1 3205 > 2 11099 > 3 11249 > 4 3317 > 5 1604
I think people are looking at the wrong thing. The purpose of these randomising routines is to ensure that we don't have collisions, NOT to be statistically 'perfect'. Does it meet that need. If yes, then don't worry about the imperfections. If no, then find a new randomiser. [...]
11314 965 43_Re: [z/OS] RAND() function strange behavior16_Gregory Palgrave28_gpalgrave@STAFF.IINET.NET.AU31_Wed, 17 Jun 2009 09:09:54 +0800324_us-ascii By way of contrast, this is the result when run on a DB2 LUW (RHEL 5) 9.5 system:
# Heads Frequency 0 1033 1 4908 2 10006 3 10067 4 4986 5 1000
Which to my unskilled eye looks like a reasonable result. Running it several times shows repeatable results within the bounds of your expected outcome. [...]
12280 31 27_How to cancel UDB utility ?3_MSH16_mvsmsh@YAHOO.COM31_Tue, 16 Jun 2009 18:11:56 -0700790_utf-8 Hi , Listers,
We are running UDB 8.2 on AIX,
Are there any UDB command like Z/OS DB2 'term utility(xxx) ' ?
I try to terminate LOAD utility because of TSM shutdown , the force application (...) doesn't works .
Regards
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG North America Attendee Testimonial- "I made a number of contacts with other people having similar experiences to mine in some of our current initiatives, giving me additional resources to discuss issues and problems with." _____________________________________________________________________
12312 28 31_Re: How to cancel UDB utility ?12_Phil Gunning22_pgunning@GUNNINGTS.COM31_Wed, 17 Jun 2009 01:23:04 +0000395_- Restart the load with terminate. Pg ------Original Message------ From: MSH Sender: DB2 Data Base Discussion List To: DB2-L@WWW.IDUGDB2-L.ORG ReplyTo: DB2 Database Discussion list at IDUG Subject: [DB2-L] How to cancel UDB utility ? Sent: Jun 16, 2009 9:11 PM
Hi , Listers,
We are running UDB 8.2 on AIX,
Are there any UDB command like Z/OS DB2 'term utility(xxx) ' ? [...]
12341 33 31_Re: How to cancel UDB utility ?16_Gregory Palgrave28_gpalgrave@STAFF.IINET.NET.AU31_Wed, 17 Jun 2009 10:02:28 +0800536_utf-8
db2 load from ..... terminate into tablename
Be aware that if you were doing a nonrecoverable load, your table will be in an unusable state after this.
All this information is in the DB2 Information Centre etc, but can be hard find when you are in a hurry.
Cheers Greg
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of MSH Sent: Wednesday, 17 June 2009 9:12 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] How to cancel UDB utility ? [...]
12375 197 38_Table controlled partitioning and DPSI8_duam lee20_duam_lee@HOTMAIL.COM31_Wed, 17 Jun 2009 03:48:19 +0000637_Windows-1252
Dear all,
I am looking for all your valuable suggestions for improvements.
A table with randomising technique used to distibute the data into index controlled parts and distribute evenly. It was okay as long as data was low volume. Now we have new v8 features and trying to convert the index controlled partition to table controlled partition. As we dont get the benefits of evenly spread we decided to go for table controlled partition. The NPI which is now having all the access to programs are c1,c2,c3,c4 and c5. One of the column (C2) is data field and I want [...]
12573 231 42_Re: Table controlled partitioning and DPSI14_Grainger, Phil20_Phil.Grainger@CA.COM31_Wed, 17 Jun 2009 08:09:15 +0100399_iso-8859-1 Duam
Firstly - you are not creating a DPSI, you are creating a PARTITIONED index - because it has the partitioning columns in the key, it is NOT a secondary index
Secondly, when you say "c2 >= date1 and c2 <= date2" I assume you mean "c2 BETWEEN date1 and date2"?
This will still allow partition pruning on the partitions that ONLY contain the dates you require [...]
12805 344 43_Re: [z/OS] RAND() function strange behavior13_Michael Ebert18_mebert@AMADEUS.COM31_Wed, 17 Jun 2009 07:54:31 +0000592_ISO-8859-1 I disagree. The purpose of a RAND() routine is to return a random number, plain and simple. That means it must meet certain minimum statistical requirements. You're right that RAND() routines are difficult to write - that's why there's half a century of research behind them. Just like SORT routines, there's absolutely no requirement to invent your own. Just pick one of the proven and tested and tuned ones that have been around for decades. However, it seems there's a strong tendency to pass this of to a student "just implement a quick RAND routine for our new product". [...]
13150 27 31_Re: TRACKMOD on DB2 8/9 for zos12_Adam Baldwin22_adambaldwin@ES.IBM.COM31_Wed, 17 Jun 2009 09:59:44 +0000306_windows-1252 Darren, I don't have any stats to give you but in a non data-sharing environment the trackmod overhead is going to be insignificant. If you turn trackmod off you will impact any incremental copy processes - they will resort to a tablespace scan - and you wont be able to use CHANGELIMIT. [...]
13178 292 43_Re: [z/OS] RAND() function strange behavior14_Avram Friedman25_ibmsysprog@GEEK-SITES.COM31_Wed, 17 Jun 2009 09:57:35 +0000375_windows-1252 I think the problem is not with RAND() but with the use of ROUND() which may skew the results. Try something like casting the random number as an integer, divide by 2 and use the remainder to which will either be 0 or 1.
With round 1/2 of the values will be increased and one 1/2 will be left alone so and expected count of 1000 will change to 1500. [...]
13471 283 43_Re: [z/OS] RAND() function strange behavior10_Dave Churn19_davechurn@GMAIL.COM31_Wed, 17 Jun 2009 07:41:22 -0500546_ISO-8859-1 Listers,
Results from DB2 9 on z/OS.
I thought there might be a problem with the ROUND function too. After reading the SQL reference and thinking about it, the effect is rather small. The RAND() function generates a number between 0 and 1. Using one decimal for rounding, I could see the result skew--5 results make zero (0, 0.1, 0.2, 0.3, 0.4) and 6 results make one (0.5, 0.6, 0.7, 0.8, 0.9, 1.0). DB2 uses 16 decmal places where 5,000,000,000,000,000 results make zero and 5,000,000,000,000,001 results make one. [...]
13755 27 43_Re: [z/OS] RAND() function strange behavior12_Pete Woodman20_pete.woodman@EDS.COM31_Wed, 17 Jun 2009 13:44:36 +0000425_windows-1252 As a slight aside, there is definately something not quite right with the RAND function ... 'cos I've run the following about 10 times and frequently got R2=R3=R4 ...
SELECT RAND() AS R1, RAND() AS R2, RAND() AS R3, RAND() AS R4 FROM SYSIBM.SYSDUMMY1
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU * [...]
13783 179 43_Re: [z/OS] RAND() function strange behavior14_M. Khalid Khan24_Mohammad_Khan@BCBSIL.COM31_Wed, 17 Jun 2009 08:45:43 -0500443_us-ascii It seems that the addition of five "random" values is playing some role in this behaviour. I ran a variation of the query without additions and I have been getting consistant results. All values are generally equally likely except the two end points which are about half as likely as rest of the values. I guess that a study of Knuth's Art of Computer Programming is in order to get a better insight into these mysteries. Khalid [...]
13963 132 38_DPSI and table controlled partitioning8_duam lee20_duam_lee@HOTMAIL.COM31_Wed, 17 Jun 2009 13:57:24 +0000473_ks_c_5601-1987
Dear all,
I am looking for all your valuable suggestions for improvements.
A table with randomising technique used to distibute the data into index controlled
parts. It was okay as long as data was low volume. Now we have new
v8 features and trying to convert the index controlled partition to table controlled
partition. As we dont get the benefits of evenly spread we decided to go for table [...]
14096 76 43_Re: [z/OS] RAND() function strange behavior14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Wed, 17 Jun 2009 10:39:36 -0400402_us-ascii SUMMARY: As Pete, Khalid, and others have observed, using RAND() multiple times in the same statement seems to return values from the same seed. That's obviously sufficient to cause the problem we're experiencing. It's also interesting that DB2 LUW apparently uses a different algorithm to generate (multiple) random numbers.
Thanks to everyone for your time and thoughts on this. [...]
14173 64 24_SQL Declare for LOCATORs6_keyRho16_keyrho@GMAIL.COM31_Wed, 17 Jun 2009 12:13:32 -0400525_ISO-8859-1 Listers,
Good day !
I wish someone tells/teaches me how could I declare SQL LOCATOR(s) in Assembler Language in Stored Procedure:
In COBOL, for example, I could code like;
01 LOCATOR USAGE SQL TYPE IS RES_SET_LOCATOR VARYING.
EXEC SQL ALLOCATE CUR#1 CURSOR FOR RESULTS SET :LOCATOR END-EXEC.
thanks in advance
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU * [...]
14238 54 28_Re: SQL Declare for LOCATORs10_Todd Burch17_toddburch@MAC.COM31_Wed, 17 Jun 2009 11:40:55 -0500420_US-ASCII MYVAR SQL TYPE IS BLOB_LOCATOR 4K
as one example. See the Application Programming and SQL Guide. Search LOCATOR
Todd
On Jun 17, 2009, at 11:13 AM, keyRho wrote:
Listers,
Good day !
I wish someone tells/teaches me how could I declare SQL LOCATOR(s) in Assembler Language in Stored Procedure:
In COBOL, for example, I could code like; [...]
14293 54 85_I will be out of the office from June 17th returning June 19th after 4pm for the IPL.12_Mark Vickers26_MarkVickers@GROCERYBIZ.COM31_Wed, 17 Jun 2009 14:46:25 -0500449_US-ASCII
I will be out of the office starting 06/17/2009 and will not return until 06/22/2009.
I will respond to your message when I return. If you need immediate assistance, please contact Ted Flowers x6562, Ramona Kipfer x5721 or Scott Stritchko x5669. thank you, Mark Vickers.
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU * [...]
14348 31 44_[FLUFF] This Tshirt Intentionally Left Blank12_Adam Baldwin22_adambaldwin@ES.IBM.COM31_Thu, 18 Jun 2009 10:31:56 +0000307_windows-1252 ...it's nearly Friday.. I remember from a year or so ago that there was a discussion re the famous "This page intentionally left blank" (thanks to Andy Lankester). I've just been sent the following link for a "This T-shirt Intentionally Left Blank" tshirt. A big blue one could be good... [...]
14380 30 52_CN=Thomas Frieske/OU=IL/O=HCSC is out of the office.11_Tom Frieske25_Thomas_Frieske@BCBSIL.COM31_Thu, 18 Jun 2009 05:37:10 -0500609_us-ascii I will be out of the office starting 06/17/2009 and will not return until 06/22/2009.
**********
The information contained in this communication is confidential, private, proprietary, or otherwise privileged and is intended only for the use of the addressee. Unauthorized use, disclosure, distribution or copying is strictly prohibited and may be unlawful. If you have received this communication in error, please notify the sender immediately at (312)653-6000 in Illinois; (800)835-8699 in New Mexico; (918)560-3500 in Oklahoma; or (972)766-6900 in Texas. [...]
14411 47 33_undocumented reason code 00E400C82_JR15_jalzola@EPO.ORG31_Thu, 18 Jun 2009 13:56:38 +0000404_ISO-8859-1 Dear DB2-L members,
00E400C8 is a reason code which sometimes we suffered from at our DB2 v8.1 on z/OS dedicated to SAP applications. There is no trace of this CR in the book, so we asked to IBM. This is the answer:
"RC00E400C8 means invalid UOBL. Found a reference from another customer where the problem occurred running a COPY job with a number of tablespaces > 32767. " [...]
14459 94 37_Re: undocumented reason code 00E400C814_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Thu, 18 Jun 2009 10:20:41 -0400431_us-ascii If your SAP tablespaces are on a separate SMS pool, you might consider SET LOG SUSPEND and pack captures instead (if these copies are for DR). Just a thought.
For your application-rollback ICs, probably you need to revisit the strategy whenever you install a new SAP module. Do distinct modules have unique databases? You could stabilize your situation by breaking up your copies at the database level, if so. [...]
14554 322 48_Re: [FLUFF] This Tshirt Intentionally Left Blank14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Thu, 18 Jun 2009 10:22:43 -0400563_us-ascii I have something funnier but you have to have seen a particular movie... The Princess Bride...
http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/9f70/
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Adam Baldwin Sent: Thursday, June 18, 2009 6:32 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] [FLUFF] This Tshirt Intentionally Left Blank
...it's nearly Friday.. I remember from a year or so ago that there was a [...]
14877 84 37_Re: undocumented reason code 00E400C814_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Thu, 18 Jun 2009 10:20:41 -0400431_us-ascii If your SAP tablespaces are on a separate SMS pool, you might consider SET LOG SUSPEND and pack captures instead (if these copies are for DR). Just a thought.
For your application-rollback ICs, probably you need to revisit the strategy whenever you install a new SAP module. Do distinct modules have unique databases? You could stabilize your situation by breaking up your copies at the database level, if so. [...]
14962 220 48_Re: [FLUFF] This Tshirt Intentionally Left Blank14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Thu, 18 Jun 2009 10:22:43 -0400563_us-ascii I have something funnier but you have to have seen a particular movie... The Princess Bride...
http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/9f70/
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Adam Baldwin Sent: Thursday, June 18, 2009 6:32 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] [FLUFF] This Tshirt Intentionally Left Blank
...it's nearly Friday.. I remember from a year or so ago that there was a [...]
15183 296 68_[AD] Empower the DBA! Q-Replication Training Class, Starts July 13th12_Kimberly May28_kim.may@THEFILLMOREGROUP.COM31_Thu, 18 Jun 2009 10:35:11 -0400387_US-ASCII I know times are tough and budgets are tight. Training can't fix everything, but knowledge is power and training is the most cost effective way to provide DBAs with the skills needed to face their ever-increasing workloads. Blah blah blah - I know, but wouldn't it be nice to really understand how it works??? And wouldn't it make you more interested in working with it??? [...]
15480 244 48_Re: [FLUFF] This Tshirt Intentionally Left Blank14_Sevetson, Phil22_PSevetson@FISA.NYC.GOV31_Thu, 18 Jun 2009 10:22:43 -0400563_us-ascii I have something funnier but you have to have seen a particular movie... The Princess Bride...
http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/9f70/
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Adam Baldwin Sent: Thursday, June 18, 2009 6:32 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: [DB2-L] [FLUFF] This Tshirt Intentionally Left Blank
...it's nearly Friday.. I remember from a year or so ago that there was a [...]
15725 234 68_[AD] Empower the DBA! Q-Replication Training Class, Starts July 13th12_Kimberly May28_kim.may@THEFILLMOREGROUP.COM31_Thu, 18 Jun 2009 10:35:11 -0400387_US-ASCII I know times are tough and budgets are tight. Training can't fix everything, but knowledge is power and training is the most cost effective way to provide DBAs with the skills needed to face their ever-increasing workloads. Blah blah blah - I know, but wouldn't it be nice to really understand how it works??? And wouldn't it make you more interested in working with it??? [...]
15960 205 48_[Z/OS] Reorg Utility - Using lots of Real Memory0_37_Sangameswara.R.Panchomarthi@CHASE.COM31_Thu, 18 Jun 2009 11:32:27 -0400423_us-ascii While Monitoring the RMF reports, we have identified that some of the REORG utility jobs are using large number of Frames (1335K). Even though we do many reorgs during a day, only these two jobs are using large number of frames. Other jobs are using very less number of Frames. During our research we were unable to identify no special characteristics for the tables that are being REORGed in these two jobs. [...]
16166 19 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Thu, 18 Jun 2009 15:46:27 +0000570_Windows-1252 >Could you please let us know, if any one of you encountered this kind of situation. What could be the cause of heavy usage of Real Memory for some tables only (Not for other tables).
From the looks of your excerpted report, they're not paging. z/OS doesn't take memory away from tasks for no reason. If your applications aren't suffering from page-in delay, and your system isn't paging, then the reason for the 'heavy' usage is because the jobs referenced the memory and nobody else has needed any since. - Too busy driving to stop for gas!
16186 70 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory15_Blaicher, Chris22_Chris_Blaicher@BMC.COM31_Thu, 18 Jun 2009 11:35:18 -0500471_iso-8859-1 I can only say what our LOAD and REORG products do, but IBM may be doing a similar process.
With BMC LOADPLUS for DB2 we look at the state of the machine and the size of the work we will have to process and balance performance and resource usage.
Sorts are memory sensitive in the sense that if you give it more memory it will generally perform better. There are diminishing returns, so there are limits based on the amount of data to sort. [...]
16257 82 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory0_37_Sangameswara.R.Panchomarthi@CHASE.COM31_Thu, 18 Jun 2009 12:49:34 -0400470_iso-8859-1 These REORG Jobs ran on Sunday early morning during very less online activity. During this time, REORG jobs has stolen the unused Frames from the CICS regions (Which has reduced the CICS Frame Count).
But on Monday morning, when the online activity ramped up, CICS was unable to catch up because of the Paging that needed to take place to get the required memory frames for the Peak online activity. We missed some of the SLAs due to this paging. [...]
16340 30 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Thu, 18 Jun 2009 17:24:35 +0000528_- >These REORG Jobs ran on Sunday early morning during very less online activity. During this time, REORG jobs has stolen the unused Frames from the CICS regions (Which has reduced the CICS Frame Count).
This not an unusual phenomenon. But, there is a solution.
>But on Monday morning, when the online activity ramped up, CICS was unable to catch up because of the Paging that needed to take place to get the required memory frames for the Peak online activity. We missed some of the SLAs due to this paging. [...]
16371 60 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory0_27_Jeff.D.Jerry@WELLSFARGO.COM31_Thu, 18 Jun 2009 12:28:47 -0500385_us-ascii Beware of using LISTDEF processing on large tables, with many columns, and using in-line Runstats in the job.
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Ted MacNEIL Sent: Thursday, June 18, 2009 12:25 PM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] [Z/OS] Reorg Utility - Using lots of Real Memory [...]
16432 407 30_Re: REXX, SQL and FOR BIT DATA15_Leland Goldston23_lgoldston@EARTHLINK.NET31_Thu, 18 Jun 2009 17:56:47 +0000677_-
/* REXX */
/*trace i */ /* initialize constants, varibables, arrays, etc */
call Init_ialize
call Check_DB2_cmd_environ
call Get_DBRM_info
do ix = 1 to DBRM_info.0
call Parse_DBRM_info
say date('s') time() '> Getting SYSPACKDEP/STMT details for package', '('ix')' DBRM_pkg_name DBRM_contoken
call Process_pkg_info
end
say date('s') time() '> Total DB2_ISUD update 'ISUD_u say date('s') time() '> Total DB2_ISUD insert 'ISUD_i say date('s') time() '> Total DB2_ISUD delete 'ISUD_d [...]
16840 59 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory15_Blaicher, Chris22_Chris_Blaicher@BMC.COM31_Thu, 18 Jun 2009 13:12:29 -0500316_us-ascii I am interested in memory utilization and how the operating system responds to various demands, so I have a few follow-up questions.
Are you saying the REORG job was still running, or that the frames once stolen by the system to satisfy the needs of the REORG are slow to be given back to CICS? [...]
16900 77 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory0_37_Sangameswara.R.Panchomarthi@CHASE.COM31_Thu, 18 Jun 2009 15:46:39 -0400449_us-ascii Please check my answers below.
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Blaicher, Chris Sent: Thursday, June 18, 2009 2:12 PM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] [Z/OS] Reorg Utility - Using lots of Real Memory
I am interested in memory utilization and how the operating system responds to various demands, so I have a few follow-up questions. [...]
16978 60 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory0_37_Sangameswara.R.Panchomarthi@CHASE.COM31_Thu, 18 Jun 2009 15:48:43 -0400307_us-ascii I have checked with the System Performance team, They are resistant to the idea of defining a service class as Storage Critical (MEMORY CRITICAL). According to them, it is a bad idea to specify this parameter, unless we have very large amount of real storage to back it up.
Thanks, Rao [...]
17039 40 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Thu, 18 Jun 2009 20:21:58 +0000338_- >I have checked with the System Performance team, They are resistant to the idea of defining a service class as Storage Critical (MEMORY CRITICAL).
Well, they're entitled to their opinion.
>According to them, it is a bad idea to specify this parameter, unless we have very large amount of real storage to back it up. [...]
17080 226 48_Re: [FLUFF] This Tshirt Intentionally Left Blank11_Nick Cianci19_ncianci@AU1.IBM.COM31_Fri, 19 Jun 2009 10:14:46 +1000432_ISO-8859-1 Phil from a dad who has sat up watching the same with his daughter ... ROFL
Long may the "cliffs of insanity" stand!
Regards, Nick CIANCI
"Sevetson, Phil" To Sent by: DB2 Data DB2-L@WWW.IDUGDB2-L.ORG Base Discussion cc List Re: [DB2-L] [FLUFF] This Tshirt Intentionally Left Blank
19/06/09 12:22 AM [...]
17307 67 73_AUTO: Michael Theunissen/TMG is out of the office. (returning 2009/06/22)0_30_Michael.Theunissen@SUNLIFE.COM31_Thu, 18 Jun 2009 21:58:17 -0400426_ISO-8859-1 I am out of the office until 2009/06/22.
For DBA work, please open a request tracker item. For anything urgent, please contact the DBA's below, or call the oncall DB2 DBA through the help desk or batch ops.
For DB2 support, contact Debbie Savage.
Note: This is an automated response to your message "DB2-L Digest - 18 Jun 2009 to 19 Jun 2009 (#2009-179)" sent on 18/6/09 21:00:10. [...]
17375 49 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory16_Christian Michel18_cmichel@DE.IBM.COM31_Fri, 19 Jun 2009 07:33:37 +0000584_ISO-8859-1 The main user of memory in the REORG utility are the different DFSORT tasks that are attached by the utility. The amount of storage requested depends on the available storage, the DFSORT installation options and also the number of parallel tasks that were started to process the indexes in parallel. Our suggestion is to set the DSA installation option of DFSORT to the maximum amount of memory that you want to give to each individual DFSORT task, the total amount of memory used is then dependend on the number of parallel tasks and the available storage of course. [...]
17425 37 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory15_Thierry Deleris44_thierry.deleris_ssii@I-BP.BANQUEPOPULAIRE.FR31_Fri, 19 Jun 2009 09:36:46 +0000378_windows-1252 Hello,
We encountered the same problem next year with some DB2 V8 Utilities who use DFSORT for parallel Index constructions (REORG IX, LOAD, RECOVER IX, ...)
Our solution : Put a REGION/MEMLIMIT on the JCL with an "empiric" value like 64M : DFSORT will adapt himself and run less parralel Index sorting with an adapted Memory value for each one. [...]
17463 36 48_Re: [FLUFF] This Tshirt Intentionally Left Blank16_mocion@libero.it16_mocion@LIBERO.IT31_Fri, 19 Jun 2009 05:01:18 -0500547_ISO-8859-1 I prefere:
'This brain was intentionally left blank'.
I saw it at least once lately......and it was not printed on a T-Shirt....:-).
Max Scarpa Certified intentionally left blank DB2 sysprog Certified flying DB2 sysprog
http://www.channeldb2.com/profile/MaxScarpa52
IDUG European 2008 Attendee Testimonial
"The technical content of the conference was excellent and useful. But in Rome it'll be even better, Max Scarpa will held 'Baldness camouflage performance and tuning' seminar :-) 8-) [...]
17500 44 12_QMF Question4_Anil21_alisha_kale@YAHOO.COM31_Fri, 19 Jun 2009 13:54:00 +0000398_windows-1252 Hi All !
Question Regarding QMF V9 on z/OS:
Why would I not see objects using the LIST command (specifying correct TYPE, OWNER, and NAME), but I can view them using the DISplay command.
For example: LIST PROCS (OWNER=XYZ NAME=MY_PROC_01 No objects were found to satisfy your LIST command.
However, DIS PROC XYZ.MY_PROC_01
would display the PROC [...]
17545 24 30_Re: REXX, SQL and FOR BIT DATA15_Leland Goldston23_lgoldston@EARTHLINK.NET31_Fri, 19 Jun 2009 15:04:59 +0000647_windows-1252 I appreciated the discussion on this topic. Also, to remind eveyone that this SQL works via SPUFI, DSNTIAUL and FileAid/DB2 without any special function in the predicate.
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG North America Attendee Testimonial- "The most important thing I always find at IDUG is the revisiting of old assumptions, with new functions and features." _____________________________________________________________________
17570 48 30_Re: REXX, SQL and FOR BIT DATA15_Leland Goldston23_lgoldston@EARTHLINK.NET31_Fri, 19 Jun 2009 17:06:56 +0000423_windows-1252 Update: I hardcoded in the REXX/exec a LIKE with a known value that SPUFI returned a hit. The SQL statement worked. I tried this after re-reading several times REXX to DB2 and vice versa datatypes. I noted the certain types which need the quote (') passed to DB2 needed to be included in double quotes (") since this is a compound variable being passed to DB2. Consequently, I tried the likve variable to [...]
17619 56 52_Re: [Z/OS] Reorg Utility - Using lots of Real Memory0_37_Sangameswara.R.Panchomarthi@CHASE.COM31_Fri, 19 Jun 2009 13:10:40 -0400522_iso-8859-1 These Jobs are set to Region=0M and SORTNUM 4 (4 Parallel tasks). In the SDSF output of the job, I see 960MB memory is used above the bar.
But RMF report showing the job using 1335K frames of memory (Which is close to 5 GB Real Memory).
-----Original Message----- From: DB2 Data Base Discussion List [mailto:DB2-L@IDUGDB2-L.ORG] On Behalf Of Thierry Deleris Sent: Friday, June 19, 2009 5:37 AM To: DB2-L@WWW.IDUGDB2-L.ORG Subject: Re: [DB2-L] [Z/OS] Reorg Utility - Using lots of Real Memory [...]
17676 37 49_Compiling nad Linkeditting CICS COBOL DB2 Program16_Larry D'Agostino16_ldagosti@CSC.COM31_Fri, 19 Jun 2009 19:48:02 +0000364_windows-1252 I am attempting to establish JCL processors for CICS COBOL programs with DB2 and IMS. I am receiving the following from Link-edit:
IEW2456E 9207 SYMBOL DSNHADDR UNRESOLVED. MEMBER COULD NOT BE INCLUDED FROM THE DESIGNATED CALL LIBRARY IEW2456E 9207 SYMBOL DSNHADD2 UNRESOLVED. MEMBER COULD NOT BE INCLUDED FROM THE DESIGNATED CALL LIBRARY [...]
17714 27 43_Re: [z/OS] RAND() function strange behavior11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Fri, 19 Jun 2009 23:40:08 +0000352_- >I disagree. The purpose of a RAND() routine is to return a random number, plain and simple. >That means it must meet certain minimum statistical requirements.
Again, I disagree. The purpose of the RAND() function is to ensure that there are no collisions. NOT to meet somebody's statistical theory. If it does that, what makes it wrong? [...]
17742 32 43_Re: [z/OS] RAND() function strange behavior14_Phil Smith III15_lists@AKPHS.COM31_Fri, 19 Jun 2009 21:09:58 -0400496_US-ASCII While I suspect this horse is dead, it's worth asking where DB2 gets its entropy. This may vary depending on whether a CEX2C is installed or not; I don't know. But if it does, and one isn't, don't expect vary random results.
The distributed DB2 probably works with a system-provided source of entropy--typically some combination of timers, mouse position, process IDs, and so forth. Some of those exist on z; some don't. It's a non-trivial problem, hence the CEX2C solution. [...]
17775 24 31_Ian Gough is out of the office.9_Ian Gough15_Ian.Gough@CN.CA31_Fri, 19 Jun 2009 22:03:11 -0400672_US-ASCII I will be out of the office starting 2009/06/18 and will not return until 2009/06/23.
If this is an emergency please contact ITCC and they will page the ITOPS-SYSTEM-DB person on-call.
_____________________________________________________________________
* IDUG 09 Rome, Italy * October 5-9, 2009 * http://IDUG.ORG/EU *
_____________________________________________________________________
IDUG DB2-L FAQ and e-mail settings are located on the IDUG.org Listserv page. If you are not already an IDUG.org member, please register at http://www.idug.org/register _____________________________________________________________________
17800 65 43_Re: [z/OS] RAND() function strange behavior14_Peter Vanroose17_pvanroose@ABIS.BE31_Sat, 20 Jun 2009 12:27:04 +0000383_ISO-8859-1 Ted MacNEIL wrote: >The purpose of the RAND() function is to ensure that there are no collisions.
I must disagree here. After having selected an integer between 0 and 99, the chance of a "collision", i.e., the next number being identical to the previous one, with a uniform random number generator, should be exactly 1%, centrainly not 0. [...]
17866 34 43_Re: [z/OS] RAND() function strange behavior14_James Campbell25_jacampbell@ACSLINK.NET.AU31_Sat, 20 Jun 2009 22:51:38 +1000577_US-ASCII Just because you want to use RAND to avoid collisions does not mean that that is its purpose. That is what GENERATE_UNIQUE is for.
James Campbell
On 19 Jun 2009 at 23:40, Ted MacNEIL wrote:
> >I disagree. The purpose of a RAND() routine is to return a random number, plain and simple. >That means it must meet certain minimum statistical requirements. > > Again, I disagree. > The purpose of the RAND() function is to ensure that there are no collisions. > NOT to meet somebody's statistical theory. > If it does that, what makes it wrong? > [...]
17901 28 43_Re: [z/OS] RAND() function strange behavior11_Ted MacNEIL18_eamacneil@YAHOO.CA31_Sat, 20 Jun 2009 15:07:57 +0000335_- >Just because you want to use RAND to avoid collisions does not mean that that is its purpose. >That is what GENERATE_UNIQUE is for.
I never said I did. I just assumed that was what the original poster wanted.
I was more in line with discussing the values returned from RAND(). - Too busy driving to stop for gas! [...]
17930 62 43_Re: [z/OS] RAND() function strange behavior14_Peter Vanroose17_pvanroose@ABIS.BE31_Sat, 20 Jun 2009 19:48:23 +0000505_ISO-8859-1 Phil,
By rewriting your query such that there is at most one RAND() on a SELECT line instead of five, the statistics became correct:
WITH BINARY_SAMPLE (SAMPLE_NO, BIN) AS ( SELECT 1, INT(ROUND(RAND(),0)) FROM SYSIBM.SYSDUMMY1 UNION ALL SELECT SAMPLE_NO+1, INT(ROUND(RAND(),0)) FROM BINARY_SAMPLE WHERE SAMPLE_NO < 160000 ) SELECT BIN_SUM, COUNT(*) AS FREQUENCY FROM ( SELECT SUM(BIN) AS BIN_SUM FROM BINARY_SAMPLE GROUP BY SAMPLE_NO/5 ) BINARY_DISTRIBUTION GROUP BY BIN_SUM [...]
17993 46 59_Any way to connect to DB2 9.1 Solaris from DB2 v7.2 client?12_Peter Suhner24_peter_suhner@HOTMAIL.COM31_Sat, 20 Jun 2009 20:44:15 +0000412_windows-1252 Dear Listers,
one of our customers still runs an old application which is stuck with Java 1.1.6 (limited lifespan, unreasonable efforts to upgrade). As a result, they can't use a DB2 Java driver newer than v7.2.
Is there any way for such a Windows application to connect to DB2 9.1 (64bit, Solaris)? According to the documentation, v7 Java (Type 2) connections are not supported. [...]
18040 76 63_Re: Any way to connect to DB2 9.1 Solaris from DB2 v7.2 client?13_Ian Bjorhovde23_ian.bjorhovde@GMAIL.COM31_Sun, 21 Jun 2009 08:27:19 -0700399_ISO-8859-1 Although I think that, hypothetically, you could come up with a workaround using the Type 3 JDBC driver on the client, a DB2 V8 instance (with the the db2jd DB2 Java Daemon), the fact that the DB2 V9.1 instance is 64-bit probably makes this impossible. V7 clients were only compatible with instances that were the same bit-width (i.e. 32bit V7 client can't talk to 64bit V7 server). [...]