1 WWW.IDUGDB2-L.ORG /home/listserv/home/db2-l December 2004, week 5
2 53 45_AW: [DB2-L] Offloading SMF accounting records19_Weber, Thomas (LDS)23_Thomas.Weber@LDS.NRW.DE31_Wed, 29 Dec 2004 07:42:56 +0100564_us-ascii Hi Ludmila,
try http://www.storage.ibm.com/software/sort/mvs/tricks
Regards
Thomas
> > Hello, > We are a small shop and just migrating some of our apps to > DB2. I am the > DB2 systems girl here, and I am also on the learning curve. > We have been told to keep 5 days worth of DB2 accounting > records of SMF > type 100 and 101. > > I am encountering issues in filtering the db2 100 and 101 > records from our > smf daily dataset. > I have raw smf dataset (offloaded from Man(n)datasets) and I am just > writing a [...]
56 101 37_Re: Offloading SMF accounting records13_Mertens, Bart18_bart.mertens@CZ.NL31_Wed, 29 Dec 2004 08:05:51 +0100779_- We use this:
//SMFDUMP JOB MSGLEVEL=1 //STARTING EXEC SMFDUMP,SMF=1 //SMFDUMP PROC //* DUMPEN NAAR Z781.P001.SMFDAG EN //* DUMPEN EN LEEGMAKEN SYS1.MAN1,SYS1.MAN2 NAAR SMFDAG FILE //STEP005 EXEC PGM=IFASMFDP //DUMPIN DD DSN=SYS1.MAN&SMF,DISP=SHR //DUMPOUT DD DSN=Z781.O001.DSN1.SMFDAG,DISP=MOD //SYSPRINT DD SYSOUT=* //SYSIN DD DISP=SHR,DSN=HSYS.O002.SYSPRO.PARMLIB(SMFDMP1)
Don't have the permission to view the parameters (SYSIN) but this is what the job output says: IFA010I SMF DUMP PARAMETERS IFA010I END(2400) -- DEFAULT IFA010I START(0000) -- DEFAULT IFA010I DATE(1900000,2099366) -- DEFAULT IFA010I OUTDD(DUMPOUT,TYPE(100:102)) -- SYSIN IFA010I INDD(DUMPIN,OPTIONS(DUMP)) -- SYSIN IFA020I DUMPOUT -- Z781.O001.DSN1.SMFDAG IFA020I DUMPIN -- SYS1.MAN1 [...]
158 42 41_Re: SQL Question , V7 and V8 difference ?14_James Campbell25_jacampbell@ACSLINK.NET.AU31_Wed, 29 Dec 2004 04:48:25 -0600669_- What about "TBSPACEID is not a column in the V7 view"?
James Campbell
On Tue, 28 Dec 2004 11:30:38 -0800, Ravi Reddy wrote:
>Hello, > >DB2 UDB on AIX. > >Iam unable to run this particular query on DB2 V7 >system : > >db2 "SELECT a.tbspaceid tbl_tbsid, b.tbspaceid >ind_tbsid, count(distinct a.tabname) num_tbl, >count(distinct b.indname) num_ind FROM syscat.tables a >LEFT JOIN syscat.indexes b ON a.tabname = b.tabname >WHERE a.type='T' GROUP BY a.tbspaceid, b.tbspaceid" > >Error : SQL0206N "B.TBSPACEID" is not valid in the >context where it is used. >SQLSTATE=42703 > >The same query runs fine on a V8 system, >Iam [...]
201 28 37_Re: Offloading SMF accounting records11_Ivan Losada18_ilosada@INSAGS.COM31_Wed, 29 Dec 2004 04:58:25 -0600433_- Hi, to extract only some type of smf records from the "mans offload" we use this: //STEP1 EXEC PGM=ICEMAN //SORTLIB DD DSN=SYS1.SORTLIB,DISP=SHR //SYSOUT DD SYSOUT=* //SORTIN DD DISP=SHR, // DSN=fichero con la descarga de los man //SORTOUT DD DISP=(NEW,CATLG,DELETE), // SPACE=(CYL,(5,5)), // DSN=fichero de salida //SYSIN DD * OPTION COPY INCLUDE COND=(6,1,BI,EQ,X'66') /* //* 66 is the record type in hex (102 in this case) [...]
230 17 27_basic bp hit ratio question10_Bill Brown21_db2_dba@BELLSOUTH.NET31_Wed, 29 Dec 2004 07:12:52 -0600372_- i was just going through the process of calculating the bufferbppo hit ratio on new bp and came across a very simple question. i compute my io by adding up the randon io, sequential io, seq prefetch io, list prefetch io, and dynamic prefetch io. i then started calculating my total get pages. i addid in my randon and seq getpages and then i came upon my question. [...]
248 57 31_Re: basic bp hit ratio question35_Joel Goldstein - Responsive Systems26_joel@RESPONSIVESYSTEMS.COM31_Wed, 29 Dec 2004 09:09:48 -0500391_iso-8859-1 Bill,
The System Hit Ratio formula is: (GP-SumOfAllPagesRead)/Getpages
While we've all used hit ratios for ever, the metric you really want to track is the I/O rate/Sec. This is measurable and convertible into CPU Secs, and application delay/elapsed times.
See my article and presentation: The Buffer Pool Hit Ratio is Dead! on www.responsivesystems.com [...]
306 42 33_Re: Shark configuration questions0_30_Steve_Grimes@AISMAIL.WUSTL.EDU31_Wed, 29 Dec 2004 08:13:06 -0600313_US-ASCII Hello, Z/os 1.4, DB2 7.1 here,
Well, it looks like 8.5 Gb 3390-9's for us -- (unless a silver bullet argument is waiting out there somewhere.) Here's the response I got (from my e-mail that summarized the perspectives I got from IBM-Main and DB2-L.) At least we won't be looking at 3390-3's! [...]
349 62 38_Important: DB2-L planned server outage15_Thomas, Janis E19_Janis.Thomas@CA.COM31_Wed, 29 Dec 2004 09:13:35 -0600758_iso-8859-1 This message is to inform you of an emergency planned outage of DB2-L's server on Thursday, December 30, 2004 at 3:00 am Eastern Standard Time. This is for maintenance purposes and the outage should not last longer than 30 minutes.
Thanks, Janis Thomas Chief list owner
--------------------------------------------------------------------------------- Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at DB2-L-REQUEST@www.idugdb2-l.org. Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
412 88 37_Re: Offloading SMF accounting records12_Missy J Case28_Missy.Case@FIRSTDATACORP.COM31_Wed, 29 Dec 2004 09:30:09 -0600622_US-ASCII Lumila, Because of the large amount of SMF data we have, we use the smfstrip that only takes the DB2 records off. It can also be used to be a bit more specific in the date & or time ranges.
//EXTRACTD EXEC PGM=IFASMFDP,REGION=7192K //SYSPRINT DD SYSOUT=* //IN DD DSN=FDR.CSE933A.DB2SMF(0), // DISP=SHR //OUT1 DD DSN=TS.DBJN.SMF.SYSA4322,DISP=(NEW,CATLG,DELETE), // UNIT=TAP38,VOL=(PRIVATE,RETAIN,,30),LABEL=(1,SL,RETPD=30), // DCB=(RECFM=VBS,LRECL=32756,BLKSIZE=32760) //* //SYSIN DD * INDD(IN,OPTIONS(DUMP)) OUTDD(OUT1,TYPE(100,101,102)) DATE(2004322,2004322) SID(SYSA) START(0700) END(1600) //* //* [...]
501 69 37_Replacing SYSADM with DBADM on OS/39016_Fennell, Michael23_Michael.Fennell@FMR.COM31_Wed, 29 Dec 2004 14:59:09 -0500768_us-ascii Folks,
I just wanted to post the question - is there a down side to replacing SYSADM with DBADM. Currently we are responsible for all the databases within our subsystem, so I am not sure how much will actually change.
Mike Fennell Fidelity Investments (508) 787-9120
--------------------------------------------------------------------------------- Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at DB2-L-REQUEST@www.idugdb2-l.org. Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
571 57 41_Re: Replacing SYSADM with DBADM on OS/3909_Min Zhang34_Zhang@ISCCOMM.IDEDE.ITG.IS.IRS.GOV31_Wed, 29 Dec 2004 15:34:31 -0500450_US-ASCII Mike,
If you are using v7, I found from "DB2 UDB for OS/390 and z/OS V7 SQL Reference" DROP 5.5 has following requirement on Authorization, which I know is true when I am trying to drop Aliases. Check others SQL statements to see if SYSADM is required but not DBADM.
.......
Authorization To drop an alias, storage group, or view, the privilege set that is defined below must include at least one of the following: [...]
629 34 37_Re: Offloading SMF accounting records15_Ludmila Koganer24_ludmilakoganer@YAHOO.COM31_Wed, 29 Dec 2004 19:02:54 -0600529_- Hi Missy and others, Thanks for the reply. I am infact using the IFASMFDP. We have 3 subsystems in one LPAR. The dump program offloads all records based on the MVS SYSID. I am able to filter all Type 100 and 101 records (I didnt have 102 but now I have added.) The frustrating issue is in squeezing the records per each subsystem. I am trying to use a DFSORT step just after the SMFDP step. The input to this SORT step is the o/p produced from IFASMFDP. I am sorting on the following fields, 8 character SysID, SMF record [...]
664 75 36_Re: Placement of the Varchar Columns12_Roger Miller19_millerrl@US.IBM.COM31_Wed, 29 Dec 2004 19:14:15 -0600339_- The prefix was handled quite a while ago, so that is not a problem. As you noted, what you have is summary statements. Taking a summary statement and then presuming there are no exceptions is likely to cause some erroneous conclusions.
Roger Miller
On Mon, 27 Dec 2004 18:18:06 -0500, LL581@DAIMLERCHRYSLER.COM wrote: [...]
740 20 37_Re: Offloading SMF accounting records17_Venkat Srinivasan32_venkat_lserv@HEWITTANDLARSEN.COM31_Wed, 29 Dec 2004 21:21:05 -0600360_- Ludmila, Shouldnt you be sorting on offset 18 which is position 19 for SORT, which is where ssid is stored?. At position 15 you won't find the db2 ssid. That is why sort is not returning any o/p due to the include records condition. That should solve your issue....? Make sure that your o/p dataset is vbs and has the same lrecl as the o/p of ifasmfdp. [...]
761 14 31_Re: basic bp hit ratio question17_Venkat Srinivasan32_venkat_lserv@HEWITTANDLARSEN.COM31_Wed, 29 Dec 2004 21:24:32 -0600696_- Aside from what Joel said, you cannot use I/O counts and getpages in the same equation. The formula should compare pages read due to alll access types Vs total Getpages and not i/os done due to all pages Vs getpages. Venkat
--------------------------------------------------------------------------------- Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at DB2-L-REQUEST@www.idugdb2-l.org. Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
776 47 41_Re: Replacing SYSADM with DBADM on OS/3907_Ed Long19_rdhm99a@PRODIGY.NET31_Wed, 29 Dec 2004 19:42:51 -0800399_us-ascii Hi Mike. Happy New Year and all that stuff. Closest to the pin is not a sewing term!!
"Fennell, Michael" wrote:
Folks,
I just wanted to post the question - is there a down side to replacing SYSADM with DBADM. Currently we are responsible for all the databases within our subsystem, so I am not sure how much will actually change. [...]
824 103 37_Re: Offloading SMF accounting records12_Boone, Henry16_HBoone@GEICO.COM31_Thu, 30 Dec 2004 06:03:52 -0500736_- The following are the SORT SYSIN control statements (with comments) I use when processing SMF data:
INCLUDE COND((0015,004,CH, SMF_SID EQ, C'ASYS'), ASYS AND, (0019,004,CH, SMF_DB2ID EQ, C'DB2P'), DB2P AND, (0006,001,CH, SMF_TYPE EQ, X'64', 100 OR, 0006,001,CH, SMF_TYPE EQ, X'65', 101 OR, 0006,001,CH, SMF_TYPE EQ, X'66')) 102
HTH.
/"""""/ !_ _! (o) (o) oOOO==(_)==OOOo============================================* * * * Henry Boone * * DB2 Systems Programmer * * SOD/ESMD * * * * GEICO Insurance Companies Phone: (301) 718-5334 * * 4608 Willard Avenue Fax: (301) 986-2408 * * Chevy Chase, MD 20815 E-Mail: hboone@geico.com * * * * .oooO * *==( )==Oooo.============================================* [...]
928 20 34_Report Recover to current (Db2 v7)14_Krnic, Ivan B.18_ikmrb@ALLSTATE.COM31_Thu, 30 Dec 2004 08:55:50 -0600703_iso-8859-1 Is anyone using Report Recovery (Current) utility ? If tablespace has a STOP LRSN value of (000000000000) this utility should ignore logs created prior to current image copy. Apparently it is not.
Ivan B. Krnic DBA
--------------------------------------------------------------------------------- Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at DB2-L-REQUEST@www.idugdb2-l.org. Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
949 15 26_Installing DB2 PE Question10_Bill Brown21_db2_dba@BELLSOUTH.NET31_Thu, 30 Dec 2004 10:17:16 -0600328_- i would like to install v8.1 or 8.2 on a windows xp home edition pc. i currently do not protect my pc by requiring a password to sign on.
during the install of db2 it wants an id and password to be used for installing everything required. is there any way to bypass the entering of a password during this process? [...]
965 27 37_Replacing SYSADM with DBADM on OS/39014_Mark McCormack27_mamccormack@STATESTREET.COM31_Thu, 30 Dec 2004 12:17:59 -0500569_US-ASCII Mike,
We have one secondary authid for the whole DBA group. That authid lost sysadm authority a few years ago. It was replaced with dbadm authority for all application data bases. We have had a few minor glitches. 1. We lost the ability to -alter bufferpool. Only sysadm can do that. 2. It is necessary to have sysadm grant use of bufferpools and perhaps stogroups to the dbadm authid. 3. Dbadm can create, grant use, and drop a table where the table creator is not the dbadm's authid. DB2v7 allows dbadm to create a view where the view creator is [...]
993 92 30_Re: Installing DB2 PE Question15_Leon Katsnelson15_leon@CA.IBM.COM31_Thu, 30 Dec 2004 12:39:53 -0500514_US-ASCII On Windows DB2 runs as a Windows Service. Windows Service is background process that is not attached to a console (keyboard and screen). There are 2 ways to run a windows service. One is to run it under some user id. This is what DB2 install program is asking you for. It is not asking you for your id but for a user id it will create in Windows so that the process that it will create will have a context to run in. Obviously every id needs a password so it is asking you to provide one. You should [...]
1086 115 41_Re: DB2 on Z/OS - Enforcing List Prefetch13_Vaughan, Mike26_Vaughan.Mike@PRINCIPAL.COM31_Thu, 30 Dec 2004 11:52:32 -0600501_iso-8859-1 I'm not saying I recommend that you do this since it's really hard to say given the information provided, but there is a sort-of standard trick for converting a nested loop join into a merge-scan join by coding it as a full outer join and adding additional criteria to turn the results back into an inner-join. Note - I make no guarantees that this will be more efficient for you, and in all honesty it will probably make it worse (given the small understanding I have of your query). [...]
1202 23 34_DB2 REORG INDEX not using prefetch12_Charles Tsao21_tsao.cct@FORCES.GC.CA31_Thu, 30 Dec 2004 13:49:26 -0600447_- Fellow DBAs,
We have a 60M rows NPI with 150M+ pseudo delete entries. When I try a REORG INDEX .. SHRLEVEL CHANGE against it. DB2 uses SYNC I/O during the UNLOAD phase. On a different NPI of similar size but with fewer Pseudo delete entries, it uses prefetch.
My guess is because of the high ratio of pseudo deletes and fragmented index pages. Can someone confirm this and is there anyway I can influence DB2 to use prefetch? [...]
1226 111 30_Re: Installing DB2 PE Question12_Myron Miller22_myronwmiller@YAHOO.COM31_Thu, 30 Dec 2004 18:39:18 -0800313_us-ascii Leon and others, I've been using LocalSystem on Db2 Connect systems since early v6 with no problems. I just install and then go into manage and services and click on the DB2 services and properties/logon and change it to localsystem. After that everything works fine. Maybe this is another "trick". [...]
1338 36 25_FW: SQL Question/Help....17_Charles Valentine15_VALENCH@AOL.COM29_Thu, 30 Dec 2004 23:55:05 EST488_US-ASCII
--------------------------------------------------------------------------------- Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at DB2-L-REQUEST@www.idugdb2-l.org. Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
1375 165 25_SQL Question (re-sent)...17_Charles Valentine15_VALENCH@AOL.COM29_Fri, 31 Dec 2004 00:04:54 EST602_US-ASCII > Dear Listserv Group, 12/30/04
> Any recommendations for my friendly DBA? > Thank You & Happy New Year To All !!! > Charles > -----Original Message----- > From: Polovsky, Richard > Sent: Tuesday, December 28, 2004 1:54 PM > Subject: SQL Question > Charles, > > This is what I'm trying to do. > The single select will be created as a view > I can't do this with cursor processing. > Let me know if you have any ideas > > Thanks > Richard D. Polovsky > ---------------------------------------------------------------------- > table |------KEY----------| > > mkt-d pri_plan DATA > > [...]
1541 34 35_Quiescing the catalog and directory10_Alan Smith21_smithac@NILDRAM.CO.UK31_Fri, 31 Dec 2004 08:05:43 +0000390_US-ASCII The DB2 utility guide has the following advice on quiecing the catalog and directory:
If a point-in-time recovery is planned for the catalog and directory, a separate QUIESCE control statement for DSNDB06.SYSCOPY is required after you quiesce the other catalog and directory table spaces. Recommendation: Recover all catalog and directory objects to the current state. [...]
1576 77 38_Re: DB2 REORG INDEX not using prefetch13_Michael Ebert18_mebert@AMADEUS.NET31_Fri, 31 Dec 2004 09:25:01 +0100587_US-ASCII I'm not sure whether the situation you describe is possible. How do you know you've got 150M+ pseudo-deleted entries?
I've rarely used REORG INDEX SHRLEVEL CHANGE, but if I remember correctly, DB2 does an unload in sorted order - it does not do a sequential scan followed by a sort (it also expands all RID lists, stores the data in the unload file, and then reverses the whole process in the reload phase - dramatic improvement possibilities for I/O and CPU are obvious). So I'd say your index is simply highly fragmented. Remember, an index as a B-tree variant is [...]
1654 80 39_Re: Quiescing the catalog and directory13_Michael Ebert18_mebert@AMADEUS.NET31_Fri, 31 Dec 2004 14:33:38 +0100565_US-ASCII I think the point is that if you quiesce SYSCOPY along with the other TSs, and then recover to that PIT, then the QUIESCEs won't show up in SYSCOPY because it it updated only after the successful quiesce. If you plan to do a PIT recovery of the Cat/Dir, then you should be very sure that there is no other update activity against the Cat/Dir (in which case it does not matter which RBA you use, apart from the mentioned difference). That is, no DDL and no running utilities (apart from your QUIESCE). Otherwise you're inviting serious trouble anyway. [...]
1735 84 41_DB2 v7.1 ZOSV1R2 SQL Performance question13_arlen stovall24_arlenstovall@HOTMAIL.COM31_Fri, 31 Dec 2004 14:16:21 +0000657_- I have question about the performance of two SQL statements. The table in questions has 200 million records currently. The buffer pools are not large enough to hold the whole index. Here is the first SQL statement:
SELECT SGJTAMRM ,SGJTRIDX,OWNERID, SGJEVTYP, SGJBATCH,SGJITEM, SGJTRIDX, SGJTYTYP FROM DISETS.WEA2TRAN WHERE OWNERID = '18045' AND SGJEVTYP = 'ADJ' AND SGJBATCH = '0000000025' AND SGJITEM = '474' AND SGJTYTYP = 'CHRG' AND SGJTRIDX < 180000 AND SGJTRIDX = (SELECT MAX(SGJTRIDX) FROM DISETS.WEA2TRAN WHERE OWNERID = '18045' ) AND SGJEVTYP = 'ADJ' AND SGJBATCH = '0000000025' AND SGJITEM = '474' AND SGJTYTYP = 'CHRG' AND SGJTRIDX [...]
1820 297 29_Re: SQL Question (re-sent)...16_Steen, Mark, ISD26_MSteen414@WORLDSAVINGS.COM31_Fri, 31 Dec 2004 08:19:05 -0600411_us-ascii on UDB, it's simple, so I'm guessing your talking Z/OS
SELECT CASE WHEN (mkt-id = 'CPT' or mkt-id ='RNY') and pri-plan = 'PLANA' THEN 'RNY' WHEN (mkt-id = 'CPT' or mkt-id ='RNY') and pri-plan = 'PLANB' THEN 'CPT' WHEN (mkt-id = 'CPT' or mkt-id ='RNY') and pri-plan = 'PLANC' THEN 'RNY' WHEN (mkt-id = 'CPT' or mkt-id ='RNY') and pri-plan = 'PLANC' THEN 'RNY' ELSE mkt-id END CASE, pri_plan [...]
2118 92 45_Re: DB2 v7.1 ZOSV1R2 SQL Performance question14_Grainger, Phil20_Phil.Grainger@CA.COM31_Fri, 31 Dec 2004 14:42:09 -0000337_utf-8 It might be me (nearly time foi another long weekend after all) but I am struggling to make logical sense of the WHERE clause!
Ignoring the fact that there is one too many ")" (I asume the one on the line WHERE OWNERID = '18045' ) shouldn't be there, but the one on line AND SGJTRIDX = 180000 ) should for starters...) [...]
2211 159 45_Re: DB2 v7.1 ZOSV1R2 SQL Performance question13_arlen stovall24_arlenstovall@HOTMAIL.COM31_Fri, 31 Dec 2004 14:55:00 +0000599_- Phil, you are right. Cut and paste erorr The SQL should have been like this.
SELECT SGJTAMRM ,SGJTRIDX INTO :WS-SGJTAMRM ,:WS-SGJTRIDX FROM WEA2TRAN WHERE OWNERID = :WEA2TRAN-SGJOWN-ID AND SGJEVTYP = :WEA2TRAN-SGJEVTYP AND SGJBATCH = :WEA2TRAN-SGJBATCH AND SGJITEM = :WEA2TRAN-SGJITEM AND SGJTYTYP = 'ALOC' AND SGJTRIDX < :WEA2TRAN-SGJTRIDX AND SGJTRIDX = (SELECT MAX(SGJTRIDX) FROM WEA2TRAN WHERE OWNERID = :WEA2TRAN-SGJOWN-ID AND SGJEVTYP = :WEA2TRAN-SGJEVTYP AND SGJBATCH = :WEA2TRAN-SGJBATCH AND SGJITEM = :WEA2TRAN-SGJITEM AND SGJTYTYP = 'ALOC' AND SGJTRIDX < :WEA2TRAN-SGJTRIDX) [...]
2371 49 25_Active log dataset rename13_Bell, Raymond22_raymond.bell@LANDG.COM31_Fri, 31 Dec 2004 14:58:57 -0000458_iso-8859-1 Colleagues,
A brain teaser before I sod off for NY (that's New Years', not New York. Chance would be a fine thing...) to do with renaming active log datasets.
Ours are named badly and inconsistently. One or the other I can - almost - live with, but both is unacceptable. Net result is that for some of our subsystems I'm going to be adding a completely new set of archive log datasets and completely removing the existing ones. [...]
2421 124 29_Re: Active log dataset rename13_Michael Ebert18_mebert@AMADEUS.NET31_Fri, 31 Dec 2004 16:12:10 +0100582_US-ASCII Stop DB2, print BSDS contents, rename VSAM files for the active logs (or repro REUSE [or ADRDSSU COPY without DEL] to predefined logs with the proper name in case you want to be able to fall back - do not pre-format, that's useful only for adding new empty logs), delete old logs from BSDS, define new logs with the proper start/end-RBAs, print BSDS again for verification, start DB2. Maybe write a REXX to create the proper control statements; that's more reliable. Where did you get that "trick" from? I once deleted/defined logs so that they'd be used in the order [...]
2546 161 29_Re: Active log dataset rename13_Bell, Raymond22_raymond.bell@LANDG.COM31_Fri, 31 Dec 2004 15:31:44 -0000545_iso-8859-1 Hey Michael,
You know, I didn't even think about renaming the logs. That's a much better idea, as long as I get the RBA ranges right. Which I will, of course...
That 'trick' I worked out while trying to get a subsystem to grind to a halt awhile ago by removing active log datasets from the BSDS. Basically I wanted the archive offload to fail. When putting the logs back I noticed DB2 switches to the log that's at the 'top' of the queue, as listed by DSNJU004. So to get them used in the 'right' order you have to [...]