Prev: DB2 V9.1 Hadr Delay Syncup
Next: db2 hangs with diaglevel 4 on startup if log files are deleted
From: danfan on 10 Jun 2010 06:12
Hi!

There are connections to PRODB - yes, but my culprit is TESTDB and I
cant see any connections to it.
There is no HADR.
ps -fL returns process from the current login only.
ps -efL returns all db2sys processes.

However, at the second try, restoring the database from an old cold
backup the "flag" was reset
and I could finally drop the database and perform a proper
incremental restore from PRODDB to TESTDB.
My guess is that "restore" had some flag set (in one or more
partitions) indicating that the restore was still in progress although
the database was online and fully available.
Question: Must "db2 restore db incremental ABORT" be run with db2_all
in a partitioned database?

Thanks for your help!
/dg


On 10 Juni, 08:15, Helmut Tessarek <tessa...(a)evermeet.cx> wrote:
> As you can see from the db2pd output, there are connections to the database.
> Therefore the database is in use.
>
> ps -fL should give you all processes and threads from the instance user. Where
> are the db2sysc processes/threads?
>
> You said db2 list applications did not return any applications.
> How can it be that there are agents connected to the database then?
>
> I also asked, if HADR was running. Is it running?
>
> On 10.06.10 1:46 , danfan wrote:
>

>
> --
> Helmut K. C. Tessarek
> DB2 Performance and Development
>
> /*
>    Thou shalt not follow the NULL pointer for chaos and madness
>    await thee at its end.
> */

From: danfan on 10 Jun 2010 06:17
On 10 Juni, 11:26, Ian <ianb...(a)mobileaudio.com> wrote:
> On Jun 9, 2:23 pm, danfan46 <danfa...(a)hotmail.com> wrote:
>
>
>
> > Hi!
>
> > (db2 UDB V9.5.5 Linux RedHat 4)
>
> > I want to drop a database before doing a restore into it from another database.
> > db2 get db cfg:
> > Database is consistent   YES
> > Rollforward pending      NO
> > Restore pending          NO
> > Automatic maintenance    ON
>
> > Deactivate database --> Success but database was not activated
>
> > Quiesce database .... --> Database is already quiesced
> > list active databases  --> not listed
> > list applications      --> not listed
>
> > Drop database .... --> SQL1371W The database is currently in use SQLSTATE=57019
>
> You can't drop a quiesced database.  You have to unquiesce it first,
> then drop it.
>
> Ian

Ah, thats it!
Thank you!
/dg
From: Helmut Tessarek on 10 Jun 2010 14:58
> ps -fL returns process from the current login only.

I know. I wanted to see the processes/threads from the instance user.
I assumed that you are logging in with the instance user, since you ran db2pd
-edus as well.

> ps -efL returns all db2sys processes.

Yep, but it actually returns all processes and threads on the system and I
wanted to avoid a long output.

> My guess is that "restore" had some flag set (in one or more
> partitions) indicating that the restore was still in progress although
> the database was online and fully available.

This flag was the quiesce flag (as Ian mentioned).
The reason why it probably worked was because you did a db2stop and db2start
at some time.
db2stop/db2start removes this 'flag' as well.

> Question: Must "db2 restore db incremental ABORT" be run with db2_all
> in a partitioned database?

This depends on how you started the restore.

--
Helmut K. C. Tessarek
DB2 Performance and Development

/*
Thou shalt not follow the NULL pointer for chaos and madness
await thee at its end.
*/
First  |  Prev  | 
Pages: 1 2
Prev: DB2 V9.1 Hadr Delay Syncup
Next: db2 hangs with diaglevel 4 on startup if log files are deleted