DSN9023I on -START on DB2 4.1
Birk, Tim
We finally got DB2 upgraded to 4.1 and are very excited about it. My
enthusiasm has been tempered somewhat however by the phone calls I've
been getting at night. I'm hoping someone can help me with my problem.
We load dozens of tables very night. Our QMF access is also available
24 hours a day. To avoid conflicts each load step is preceded with a
step that starts the tablespace access(UT) followed by a display of
users. If there are any, a console display lists them and waits for an
operator reply when they've all been eliminated.
Our problem is with the -START. Occasionally it abnormally terminates
with message DSN9023I. There are no other messages in the job or in the
DB2 log to indicate why the problem occurred. The -DIS still runs and
shows that the tablespace is still RW. The cond code is set to 12 which
prevents other steps in the job from running. When we rerun the job,
it's fine. This did not happen on 3.1. Does anyone have any idea why
this is happening or what I can do to prevent it?
Following is a sample:
ACF0C038 ACF2 LOGONID ATTRIBUTES HAVE REPLACED DEFAULT USER ATTRIBUTES
READY
DSN SYSTEM(DSN)
DSN
-TERM UTIL(P418183D.LOAD3)
DSNU112I - DSNUGDIS - NO AUTHORIZED UTILITY FOUND FOR UTILID =
P418183D.LOAD3
DSN9022I - DSNUGCCC '-TERM UTIL' NORMAL COMPLETION
DSN
-STA DB(DBCCEF) SPACE(CCEFPROD) ACCESS(UT)
DSN9023I - DSNTDDIS 'START DATABASE' ABNORMAL COMPLETION
DSN
-DIS DB(DBCCEF) SPACE(CCEFPROD) USE
DSNT360I - ***********************************
DSNT361I - * DISPLAY DATABASE SUMMARY
* GLOBAL USE
DSNT360I - ***********************************
DSNT362I - DATABASE = DBCCEF STATUS = RW
DBD LENGTH = 20180
DSNT397I -
NAME TYPE PART STATUS CONNID CORRID USERID
-------- ---- ---- ------------------ -------- ------------ --------
CCEFPROD TS RW
******* DISPLAY OF DATABASE DBCCEF ENDED **********************
DSN9022I - DSNTDDIS 'DISPLAY DATABASE' NORMAL COMPLETION
DSN
END
Tim Birk
Senior DBA
Diebold, Inc.
Tim
Yassin Isaac
This is a multi-part message in MIME format.
--------------79CFB92E4B3B124BEB1B6B5F
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi,
Try first stopping the TS and than starting "UT".
Isaac Yassin
Birk, Tim wrote:
--------------79CFB92E4B3B124BEB1B6B5F
Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Isaac, Yassin
Content-Disposition: attachment; filename="vcard.vcf"
begin: vcard
fn: Yassin Isaac
n: Isaac;Yassin
org: Databases & Information Systems
adr: 9 Yanush Korchak st.;;;Hertzliya;;46302;Israel
email;internet: yas...@netvision.net.il
tel;work: 972 52 452793
tel;fax: 972 9 9560803
tel;home: 972 9 9505172
x-mozilla-cpt: ;0
x-mozilla-html: FALSE
version: 2.1
end: vcard
--------------79CFB92E4B3B124BEB1B6B5F--
Birk, Tim
Thanks. I've checked into that and for us it's a problem. If the
tablespace is in use, issuing a -stop in batch will hang until it's no
longer in use.
> ----------
> From: Yassin Isaac[SMTP:yas...@NETVISION.NET.IL]
> Sent: Friday, November 14, 1997 12:51 PM
> To: DB...@AMERICAN.EDU
> Subject: Re: DSN9023I on -START on DB2 4.1
>
> <<File: vcard.vcf>>
Yassin Isaac
This is a multi-part message in MIME format.
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi,
Start ut will not prevent other users from joining the game . Stop will wait
till the last one has finished and than you issue the start ut. I know it
means that short term requests will not be done but at least you stop it and
get your shot at it.
A more "active" access is to analyze the users list from the display command
and terminate them.....
Isaac Yassin
Birk, Tim wrote:
--------------E29568CAADE8F7478B1006C4
Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Isaac, Yassin
Content-Disposition: attachment; filename="vcard.vcf"
begin: vcard
fn: Yassin Isaac
n: Isaac;Yassin
org: Databases & Information Systems
adr: 9 Yanush Korchak st.;;;Hertzliya;;46302;Israel
email;internet: yas...@netvision.net.il
tel;work: 972 52 452793
tel;fax: 972 9 9560803
tel;home: 972 9 9505172
x-mozilla-cpt: ;0
x-mozilla-html: FALSE
version: 2.1
end: vcard
--------------E29568CAADE8F7478B1006C4--
Mark A. Mccormack
Tim,
1. Msg DSN9023I is not very informative. Have you checked the DB2MSTR log to
see if some other, more informative msgs were written there ? Not all DSN---
msgs are written to SYSPRINT in batch steps.
2. I also was in the habit of issuing -START DB(-) SPACE(-) ACC(UT) prior to
running utilities like load and reorg. I have had enough problems with that
approach that I have needed to take actions to protect myself.
a. Any status-changing command can cause the tablespace in question to go into
START or STOP PENDING status if it is in use by some other process, such as
QMF.
b. If you take actions to insure there are no users or claimers before issuing
the -START command, you probably can avoid START PENDING, and I suspect that
START PENDING is at the root of your problem.
c. It sounds as if you are issuing -START .. ACC(UT) and -DIS .. USE from a
program. Could you reverse the order of the commands, delaying the -START
command until the response to the WTOR indicates that the coast is clear ?
3. Here's what I have done.
a. Issue -DIS DB(-) SPACE(-) USE and -DIS DB(-) SPACE(-) CLAIMERS
b. Analyze the reports with a small home-grown program.
b1. If the coast is not clear, the program issues rc=8. The -DIS cmd reports
are also written out. The job stops. It will be restarted later.
b2. If the coast is clear, the job continues.
c. Issue -START .. ACC(UT) only after is is certain that it will succeed.
d. Run the load utility.
e. Issue -START .. ACC(RW).
I have come to realize that steps c. and e. are really not needed after you
take the actions in step b.
.. The load utility will put the tablespace into UTUT status while the utility
runs, thereby protecting itself.
.. -START DB(-) SPACE(-) ACC(UT) is really not effective in forcing a
standalone window for the utility.
I am, therefore, phasing out my use of -START DB(-) SPACE(-) ACC(UT) in this
context.
Mark McCormack
BellSouth Telecommunications
standard disclaimer: my opinions are merely my own