part2: recover postponed thread, retain locks (DB2 for OS390 V6.1)

Kadir Guray Meric

part2: recover postponed thread, retain locks (DB2 for OS390 V6.1)
Hi to everyone in the list,

This is part2, the continuation of the part 1: recover postponed thread, retain locks (DB2 for OS390 V6.1):
......continued.....

We see that DB1T data sharing memeber is in AI state. Output of /F DB1TIRLM,STATUS statement is:
DXR101I DJ1T001 STATUS SCOPE=GLOBAL
SUBSYSTEMS IDENTIFIED
NAME STATUS UNITS HELD WAITING RET_LKS
DB1T UP 6 689 0 3
We see that there are 3 retain locks on member DB1T.
Altough we have LBACKOUT=AUOTO in DSNZPARM, it could not backout.
+DB1TDIS THREAD(*) TYPE(POSTPONED)
DSNV401I +DB1T DISPLAY THREAD REPORT FOLLOWS -
DSNV431I +DB1T POSTPONED ABORT THREADS - 591
COORDINATOR STATUS RESET URID AUTHID
BATCH ABORT-P 002AB444A3B2 SDBA1
We tried to "recover postponed" command.
+DB1TRECOVER POSTPONED
DSNV435I +DB1T RESOLUTION OF POSTPONED ABORT URS HAS BEEN SCHEDULED
DSN9022I +DB1T DSNVRP 'RECOVER POSTPONED' NORMAL COMPLETION
As you see, it says "URS HAS BEEN SCHEDULED" but it does not do anything. It is supposed that this command backouts for pagesets on our tablespaces with "DSNI024I -DSNIAPRL BACKOUT PROCESSING HAS COMPLETED FOR PAGESET DDTUK00.SDTUKG25 part (n/a)..." message prefix. But this message never occured. This command does nothing in our environment.
We see that -recover postponed command did not work. We have even tried to remove tablespace but it gets error,REASON=00C200EB. We know that only datasharing member DB1T has retain locks on this tablespace. We tried to stop all of 3 subsystems. DB2T and DB3T closed with stop db2 command, but DB1T did not stop. We had to use "/f db1tirlm,abend,dump" .Then we started all data sharing mambers again. But it again had 3 retain locks on member DB1T.
(We have even tried to recover this tablespace to LRSN with RECOVER TABLESPACE DDTUK00.SDTUKG25 TOLOGPOINT X'B8A856668BC5', but it failed. We can neither drop nor do someting on this tablespace.)
What is the reason for the command "-recover postponed" not to work?How can we resolve postponed threads from our data sharing member DB1T?

Thanks for everyone's comments.
Kadir Guray MERIC
Data Management Specialist

Mike Turner

Re: part2: recover postponed thread, retain locks (DB2 for OS390 V6.1)
(in response to Kadir Guray Meric)
Hi Kadir

How long did you wait for the rollback to complete? I use a (very) rough
estimate that rollback takes 3 x original run time. So if the batch job had
been running 3 hours, you might need to wait up to 9 hours for the rollback
to complete.

Regards
Mike Turner
Email: [login to unmask email]
Web: www.michael-turner.ltd.uk

Message text written by DB2 Data Base Discussion List
>Hi to everyone in the list,=20

This is part2, the continuation of the part 1: recover postponed thread, =
retain locks (DB2 for OS390 V6.1):=20
......continued.....
=20
We see that DB1T data sharing memeber is in AI state. Output of /F =
DB1TIRLM,STATUS statement is:
DXR101I DJ1T001 STATUS SCOPE=3DGLOBAL=20
SUBSYSTEMS IDENTIFIED =20
NAME STATUS UNITS HELD WAITING RET_LKS =20
DB1T UP 6 689 0 3 =20
We see that there are 3 retain locks on member DB1T.
Altough we have LBACKOUT=3DAUOTO in DSNZPARM, it could not backout.
+DB1TDIS THREAD(*) TYPE(POSTPONED) =20
DSNV401I +DB1T DISPLAY THREAD REPORT FOLLOWS - =20
DSNV431I +DB1T POSTPONED ABORT THREADS - 591 =20
COORDINATOR STATUS RESET URID AUTHID =20
BATCH ABORT-P 002AB444A3B2 SDBA1 =20
We tried to "recover postponed" command.
+DB1TRECOVER POSTPONED =20
DSNV435I +DB1T RESOLUTION OF POSTPONED ABORT URS HAS BEEN SCHEDULED=20
DSN9022I +DB1T DSNVRP 'RECOVER POSTPONED' NORMAL COMPLETION =20
As you see, it says "URS HAS BEEN SCHEDULED" but it does not do =
anything. It is supposed that this command backouts for pagesets on our =
tablespaces with "DSNI024I -DSNIAPRL BACKOUT PROCESSING HAS COMPLETED =
FOR PAGESET DDTUK00.SDTUKG25 part (n/a)..." message prefix. But this =
message never occured. This command does nothing in our environment.
We see that -recover postponed command did not work. We have even tried =
to remove tablespace but it gets error,REASON=3D00C200EB. We know that =
only datasharing member DB1T has retain locks on this tablespace. We =
tried to stop all of 3 subsystems. DB2T and DB3T closed with stop db2 =
command, but DB1T did not stop. We had to use "/f db1tirlm,abend,dump" =
.Then we started all data sharing mambers again. But it again had 3 =
retain locks on member DB1T.
(We have even tried to recover this tablespace to LRSN with RECOVER =
TABLESPACE DDTUK00.SDTUKG25 TOLOGPOINT X'B8A856668BC5', but it failed. =
We can neither drop nor do someting on this tablespace.)
What is the reason for the command "-recover postponed" not to work?How =
can we resolve postponed threads from our data sharing member DB1T?=20
=20
Thanks for everyone's comments.
Kadir Guray MERIC
Data Management Specialist
<