A strange behavior of DB2 when stopped. Any suggestion ?

Max Scarpa

A strange behavior of DB2 when stopped. Any suggestion ?
I wrote a mail some months ago where I explained this problem:

After a -STOP DB2 command you receive the message in SDSF log that is all is
ok, DB2 is

stopped and if you try run a query, for instance, you obtain the messsage that
DB2 is

not active.

But if you look at SDSF panel after a DA command to have a snapshot of all
subsystems

running you still have DB2 address spaces (MSTR, DBM1, DIST) on the panel, all
not

highlighted and not running.

The only address space missing is IRLM, stopped correctly. To get rid off these
'ghosts' you

have to use cancel command (/C DB2xMSTR, etc).

We have all PTF that IBM sent us to try to solve the problem, but it still
remains.

I checked member IEFSSN00 in S/390 PARMLIB but it's correct.

It seems that there's some interaction between DB2 and BOOLE & BABBAGE MV
monitors

during DB2 shutdown but it's only a feeling.

Has someone the same problem and/or any suggestion ? It's not a great problem
(DB2 is

running 24x24x7 so it's is stopped very infrequently) but it hurts me :-((.

Regards

Max Scarpa
Data & System Admin
Cesve SpA

**** Standard disclaimers apply ****



Max Scarpa

FW: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Max Scarpa)
Sorry, I've forgotten to say that there's absolutely no activity on DB2 and so
it's no waiting for

any terminating thread, as you can see at restart (inflight/indoubt/incommit/
inabort count = 0).

The shutdown is correct, as far as I can see from messages, at the end and at
restart of DB2.

Now I cancel DB2 after few minutes, but at the beginning I waited for a long
time (> 10 minutes)

because I was afraid to corrupt catalog.

I think it may be a problem on our S/390 customization.

Regards

Max Scarpa



Tom - Cendant Kennedy

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Max Scarpa)
After you issue the -STO DB2 command, or just before if you like, issue a
-DIS THREAD(*) command to see what connections are still active. You'll
probably see your Boole started task still connected. If you're bringing DB2
down normally it will wait for all connections to terminate. If you don't
manually stop the Boole started task it will prevent DB2 from coming down.

I'm not sure if there is a way to break the connection between Boole and DB2
without bringing Boole down. That would be cleaner if you'd like to keep
Boole up.


Tom Kennedy - Cendant
Tech Services DB2




-----Original Message-----
From: Massimo Scarpa [mailto:[login to unmask email]
Sent: Friday, January 07, 2000 9:01 AM
To: [login to unmask email]
Subject: A strange behavior of DB2 when stopped. Any suggestion ?


I wrote a mail some months ago where I explained this problem:

After a -STOP DB2 command you receive the message in SDSF log that is all
is
ok, DB2 is

stopped and if you try run a query, for instance, you obtain the messsage
that
DB2 is

not active.

But if you look at SDSF panel after a DA command to have a snapshot of all
subsystems

running you still have DB2 address spaces (MSTR, DBM1, DIST) on the panel,
all
not

highlighted and not running.

The only address space missing is IRLM, stopped correctly. To get rid off
these
'ghosts' you

have to use cancel command (/C DB2xMSTR, etc).

We have all PTF that IBM sent us to try to solve the problem, but it still
remains.

I checked member IEFSSN00 in S/390 PARMLIB but it's correct.

It seems that there's some interaction between DB2 and BOOLE & BABBAGE MV
monitors

during DB2 shutdown but it's only a feeling.

Has someone the same problem and/or any suggestion ? It's not a great
problem
(DB2 is

running 24x24x7 so it's is stopped very infrequently) but it hurts me :-((.

Regards

Max Scarpa
Data & System Admin
Cesve SpA

**** Standard disclaimers apply ****





[login to unmask email]

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Tom - Cendant Kennedy)
Max,

A possiblity is that your subsystem swapped out on shut down. If you have
MVS Omegamon, try swapping
the tasks in to see if the terminate noramlly. I would do DB2xDIST,
DB2xDBM1, then DB2xMSTR. If this works,
you need to talk to your MVS people about the performance settings of your
DB2 address spaces

Bob Pingston
Amway



Mayank Sharma

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Bob_Pingston@AMWAY.COM)
Hi

I think problem is in your definition of System to vtam.
Please Check your entry in Vtamlst

Mayank



Jeffery A Price

Re: Strange Behavior of DB2 when Stopped
(in response to Mayank Sharma)
Max,

We are DB2 V5.1 PUT9910, and have Boole and Babbage (I'm not sure which
version), and have no problems stopping DB2 normally. We also didn't have any
problems stopping DB2 when we were are maintenance level PUT9801 (had the Boole
product then too).

Best of Luck,
Jeffery A. Price
Associate Engineer
Indianapolis Power and Light Co.
[login to unmask email]



Phil Grainger

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Jeffery A Price)
An alternative, which is 'guaranteed' to work is to issue a -STOP DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for a good
reason, then presumeably you have checked that there are no active threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue any more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International


> -----Original Message-----
> From: Kennedy, Tom - Cendant [SMTP:[login to unmask email]
> Sent: Friday, January 07, 2000 3:22 PM
> To: [login to unmask email]
> Subject: Re: A strange behavior of DB2 when stopped. Any suggestion ?
>
> After you issue the -STO DB2 command, or just before if you like, issue a
> -DIS THREAD(*) command to see what connections are still active. You'll
> probably see your Boole started task still connected. If you're bringing
> DB2 down normally it will wait for all connections to terminate. If you
> don't manually stop the Boole started task it will prevent DB2 from coming
> down.
>
> I'm not sure if there is a way to break the connection between Boole and
> DB2 without bringing Boole down. That would be cleaner if you'd like to
> keep Boole up.
>
>
> Tom Kennedy - Cendant
> Tech Services DB2
>
>
>
>
> -----Original Message-----
> From: Massimo Scarpa [ <mailto:[login to unmask email]>]
> Sent: Friday, January 07, 2000 9:01 AM
> To: [login to unmask email]
> Subject: A strange behavior of DB2 when stopped. Any suggestion ?
>
>
> I wrote a mail some months ago where I explained this problem:
>
> After a -STOP DB2 command you receive the message in SDSF log that is
> all is
> ok, DB2 is
>
> stopped and if you try run a query, for instance, you obtain the messsage
> that
> DB2 is
>
> not active.
>
> But if you look at SDSF panel after a DA command to have a snapshot of
> all
> subsystems
>
> running you still have DB2 address spaces (MSTR, DBM1, DIST) on the
> panel, all
> not
>
> highlighted and not running.
>
> The only address space missing is IRLM, stopped correctly. To get rid off
> these
> 'ghosts' you
>
> have to use cancel command (/C DB2xMSTR, etc).
>
> We have all PTF that IBM sent us to try to solve the problem, but it still
>
> remains.
>
> I checked member IEFSSN00 in S/390 PARMLIB but it's correct.
>
> It seems that there's some interaction between DB2 and BOOLE & BABBAGE MV
> monitors
>
> during DB2 shutdown but it's only a feeling.
>
> Has someone the same problem and/or any suggestion ? It's not a great
> problem
> (DB2 is
>
> running 24x24x7 so it's is stopped very infrequently) but it hurts me
> :-((.
>
> Regards
>
> Max Scarpa
> Data & System Admin
> Cesve SpA
>
> **** Standard disclaimers apply ****
>
>
>
> the DB2-L webpage at < http://www.ryci.com/db2-l > . The owners of the list
> can
>



Alexander Andrade

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Phil Grainger)
I had this happen to me when I was a sysprog, when you stopped certain
subsystems like DB2 and HSM as well, it seemed to take up to 1/2hr for the
address spaces to "bugger off".

If it is happening with HSM as well then it may be SMF cutting all of its
records for every single dataset used during the life of the address space.
To stop this there is a parmlib member SMFPRMxx which has a parameter
DDCONS, if this is set to YES(the default...... I think) then an SMF record
will be cut for each and every dataset, you may want to "encourage" mr
sysprog to set DDCONS to NO.

It may be this is happening to you,........ well your db2 anyway.

Alex Andrade

Email : [login to unmask email] <mailto:[login to unmask email]>

-----Original Message-----
From: Grainger, Phil [SMTP:[login to unmask email]
Sent: Monday, January 10, 2000 1:16 PM
To: [login to unmask email]
Subject: Re: A strange behavior of DB2 when stopped. Any
suggestion ?

An alternative, which is 'guaranteed' to work is to issue a -STOP
DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for
a good
reason, then presumeably you have checked that there are no active
threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue
any more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would
rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International


> -----Original Message-----
> From: Kennedy, Tom - Cendant [SMTP:[login to unmask email]
> Sent: Friday, January 07, 2000 3:22 PM
> To: [login to unmask email]
> Subject: Re: A strange behavior of DB2 when stopped. Any
suggestion ?
>
> After you issue the -STO DB2 command, or just before if you like,
issue a
> -DIS THREAD(*) command to see what connections are still active.
You'll
> probably see your Boole started task still connected. If you're
bringing
> DB2 down normally it will wait for all connections to terminate.
If you
> don't manually stop the Boole started task it will prevent DB2
from coming
> down.
>
> I'm not sure if there is a way to break the connection between
Boole and
> DB2 without bringing Boole down. That would be cleaner if you'd
like to
> keep Boole up.
>
>
> Tom Kennedy - Cendant
> Tech Services DB2
>
>
>
>
> -----Original Message-----
> From: Massimo Scarpa [ <mailto:[login to unmask email]>]
> Sent: Friday, January 07, 2000 9:01 AM
> To: [login to unmask email]
> Subject: A strange behavior of DB2 when stopped. Any suggestion ?
>
>
> I wrote a mail some months ago where I explained this problem:
>
> After a -STOP DB2 command you receive the message in SDSF log
that is
> all is
> ok, DB2 is
>
> stopped and if you try run a query, for instance, you obtain the
messsage
> that
> DB2 is
>
> not active.
>
> But if you look at SDSF panel after a DA command to have a
snapshot of
> all
> subsystems
>
> running you still have DB2 address spaces (MSTR, DBM1, DIST) on
the
> panel, all
> not
>
> highlighted and not running.
>
> The only address space missing is IRLM, stopped correctly. To get
rid off
> these
> 'ghosts' you
>
> have to use cancel command (/C DB2xMSTR, etc).
>
> We have all PTF that IBM sent us to try to solve the problem, but
it still
>
> remains.
>
> I checked member IEFSSN00 in S/390 PARMLIB but it's correct.
>
> It seems that there's some interaction between DB2 and BOOLE &
BABBAGE MV
> monitors
>
> during DB2 shutdown but it's only a feeling.
>
> Has someone the same problem and/or any suggestion ? It's not a
great
> problem
> (DB2 is
>
> running 24x24x7 so it's is stopped very infrequently) but it hurts
me
> :-((.
>
> Regards
>
> Max Scarpa
> Data & System Admin
> Cesve SpA
>
> **** Standard disclaimers apply ****
>
>
>
visit
> the DB2-L webpage at < http://www.ryci.com/db2-l > . The owners of
the list
> can
>




can



Leslie Pendlebury-Bowe

Re: A strange behavior of DB2 when stopped. Any suggestio
(in response to Alexander Andrade)
Phil
or worse still is "cancel of the IRLM" .. :-)
which I have had to issue as the bugger would not come down with a
FORCE ...
Leslie


______________________________ Reply Separator _________________________________
Subject: Re: A strange behavior of DB2 when stopped. Any suggestion ?
Author: "Grainger; Phil" <[login to unmask email]> at Internet
Date: 1/10/00 1:16 PM


An alternative, which is 'guaranteed' to work is to issue a -STOP DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for a good
reason, then presumeably you have checked that there are no active threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue any more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International


> -----Original Message-----
> From: Kennedy, Tom - Cendant [SMTP:[login to unmask email]
> Sent: Friday, January 07, 2000 3:22 PM
> To: [login to unmask email]
> Subject: Re: A strange behavior of DB2 when stopped. Any suggestion ?
>
> After you issue the -STO DB2 command, or just before if you like, issue a
> -DIS THREAD(*) command to see what connections are still active. You'll
> probably see your Boole started task still connected. If you're bringing
> DB2 down normally it will wait for all connections to terminate. If you
> don't manually stop the Boole started task it will prevent DB2 from coming
> down.
>
> I'm not sure if there is a way to break the connection between Boole and
> DB2 without bringing Boole down. That would be cleaner if you'd like to
> keep Boole up.
>
>
> Tom Kennedy - Cendant
> Tech Services DB2
>
>
>
>
> -----Original Message-----
> From: Massimo Scarpa [ <mailto:[login to unmask email]>]
> Sent: Friday, January 07, 2000 9:01 AM
> To: [login to unmask email]
> Subject: A strange behavior of DB2 when stopped. Any suggestion ?
>
>
> I wrote a mail some months ago where I explained this problem:
>
> After a -STOP DB2 command you receive the message in SDSF log that is
> all is
> ok, DB2 is
>
> stopped and if you try run a query, for instance, you obtain the messsage
> that
> DB2 is
>
> not active.
>
> But if you look at SDSF panel after a DA command to have a snapshot of
> all
> subsystems
>
> running you still have DB2 address spaces (MSTR, DBM1, DIST) on the
> panel, all
> not
>
> highlighted and not running.
>
> The only address space missing is IRLM, stopped correctly. To get rid off
> these
> 'ghosts' you
>
> have to use cancel command (/C DB2xMSTR, etc).
>
> We have all PTF that IBM sent us to try to solve the problem, but it still
>
> remains.
>
> I checked member IEFSSN00 in S/390 PARMLIB but it's correct.
>
> It seems that there's some interaction between DB2 and BOOLE & BABBAGE MV
> monitors
>
> during DB2 shutdown but it's only a feeling.
>
> Has someone the same problem and/or any suggestion ? It's not a great
> problem
> (DB2 is
>
> running 24x24x7 so it's is stopped very infrequently) but it hurts me
> :-((.
>
> Regards
>
> Max Scarpa
> Data & System Admin
> Cesve SpA
>
> **** Standard disclaimers apply ****
>
>
>
> the DB2-L webpage at < http://www.ryci.com/db2-l > . The owners of the list
> can
>








Mike Turner

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Leslie Pendlebury-Bowe)
Phil

A -STOP DB2 MODE(QUIESCE) can be overridden by a later -STOP DB2
MODE(FORCE). I admit I haven't had to use it for a while, but it used to
work OK if the MODE(QUIESCE) was hanging because of active threads.

Regards
Mike Turner
Email: [login to unmask email]
Home Page: http://ourworld.compuserve.com/homepages/michael_turner

Message text written by DB2 Data Base Discussion List
>An alternative, which is 'guaranteed' to work is to issue a -STOP DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for a good
reason, then presumeably you have checked that there are no active threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue any
more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International<



Max Scarpa

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Mike Turner)
I'm sure, as I experimented many times, that DB2 is a robust DBMS and it can
survive

everything might happen, but I think it's not a 'normal' stopping mode using
-STOP DB2(FORCE)
every time I shutdown DB2, I'm pretty sure that DB2 will become full of dirty
things and 'impatient' .

I've experimented 2 times CRC restart and it works, but I prefere to avoid this
type of restart.

And yes Leslie, I can try with /C IRLM.....and why not with MVS command
QUIESCE.....? Just to

have a little fun :-))

Regards :-)))

Max Scarpa
Data & System Admin
Cesve SpA

Standard disclaimers apply.....



RICK (SWBT) DAVIS

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to Max Scarpa)
Hi all,
I've never used the BOOLE & BABBAGE DB2 monitor but it would seem
they should be able to capture any DB2 STOP messages, disconnect, and then
sit around waiting for DB2 to start again. Maybe a BOOLE & BABBAGE person
can answer this.
The way I see it, how DB2 is brought down DEPENDS on the reason.
Some reasons for stopping DB2 that I have seen range from "normal" to
"immediate/bizarre", each with its own causes/requirements. Aren't these the
methods for DB2 shutdown from mildest to most severe:
-STOP DB2 (implied MODE(QUIESCE) most 24x7 shops can never get DB2 down this
way without some planned global shutdown of all ASIDs accessing DB2 as well)

-STOP DB2 MODE(FORCE) which can be issued after MODE(QUIESCE) and will
normally require some recovery action during DB2 restart. Good to know
reason, if you can, before MODE(FORCE) issued.

Then MVS command P aaaaIRLM (MVS END command)

finally MVS command C aaaaIRLM (MVS cancel command, will also usually
have longer DB2 restart time and may include manual resolution for
INDOUBT/INFLIGHT threads. Can have impact of leaving pieces of fragmented
memory laying around until the next MVS IPL)

HTH,
Rick Davis
"This e-mail and any files transmitted with it are the property of SBC,
are confidential, and are intended solely for the use of the individual
or entity to whom this e-mail is addressed. If you are not one of the
named recipient(s) or otherwise have reason to believe that you have
received this message in error, please notify the sender at 314-235-6854
and delete this message immediately from your computer. Any other use,
retention, dissemination, forwarding, printing, or copying of this
e-mail is strictly prohibited."



-----Original Message-----
From: Mike Turner [mailto:[login to unmask email]
Sent: Monday, January 10, 2000 8:19 AM
To: [login to unmask email]
Subject: Re: A strange behavior of DB2 when stopped. Any suggestion ?


Phil

A -STOP DB2 MODE(QUIESCE) can be overridden by a later -STOP DB2
MODE(FORCE). I admit I haven't had to use it for a while, but it used to
work OK if the MODE(QUIESCE) was hanging because of active threads.

Regards
Mike Turner
Email: [login to unmask email]
Home Page: http://ourworld.compuserve.com/homepages/michael_turner

Message text written by DB2 Data Base Discussion List
>An alternative, which is 'guaranteed' to work is to issue a -STOP DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for a good
reason, then presumeably you have checked that there are no active threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue any
more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International<



http://www.ryci.com/db2-l. The owners of the list can be reached at
[login to unmask email]



[login to unmask email]

Re: A strange behavior of DB2 when stopped. Any suggestion ?
(in response to RICK (SWBT) DAVIS)
once you have issued the -stop db2 command, i believe you can not then create a
new thread that may be required to issue any further db2 commands (eg stop db2
mode(force)).

however i'm fairly sure you can still issue commands via the console.

steve





Mike Turner <[login to unmask email]> on 01/10/2000 09:18:49 AM

Please respond to DB2 Data Base Discussion List <[login to unmask email]>

To: [login to unmask email]
cc:
Subject: Re: A strange behavior of DB2 when stopped. Any suggestion ?


Phil

A -STOP DB2 MODE(QUIESCE) can be overridden by a later -STOP DB2
MODE(FORCE). I admit I haven't had to use it for a while, but it used to
work OK if the MODE(QUIESCE) was hanging because of active threads.

Regards
Mike Turner
Email: [login to unmask email]
Home Page: http://ourworld.compuserve.com/homepages/michael_turner

Message text written by DB2 Data Base Discussion List
>An alternative, which is 'guaranteed' to work is to issue a -STOP DB2
MODE(FORCE)

This sounds a bit nasty, but assuming you are shutting DB2 down for a good
reason, then presumeably you have checked that there are no active threads
beforehand..........

The only snag with -STOP DB2, is that once issued. you cannot issue any
more
commands (like -TERM THREAD or -STOP DB2 MODE(FORCE) so I would rather
start with the biggest hammer!!

Phil Grainger
Director DB2 Operations, Europe
Computer Associates International<