Hello Jared,
 
Thanks for sending me the file. One can see that the DBM Server has
problem to access Backint for MaxDB's output and error output after
Backint for MaxDB was killed by that same DBM Server:
 
Can not open file 'E:\sapdb\BOO\db\backups\io\backint.output'.
(System error 32; The process cannot access the file because it is being
used by another process.)
...
Copying output of Backint for SAP DB to this file.
---------- Begin of output of Backint for SAP DB
(E:\sapdb\BOO\db\backups\io\backint.output)----------
Have encountered error -24919:
Can not open file 'E:\sapdb\BOO\db\backups\io\backint.output'.
(System error 32; The process cannot access the file because it is being
used by another process.)
...
Not removing Backint for SAP DB's temporary output file
(E:\sapdb\BOO\db\backups\io\backint.output).

Copying error output of Backint for SAP DB to this file.
---------- Begin of error output of Backint for SAP DB
(E:\sapdb\BOO\db\backups\logs\backint.err)----------
Have encountered error -24919:
Can not open file 'E:\sapdb\BOO\db\backups\logs\backint.err'.
(System error 32; The process cannot access the file because it is being
used by another process.)
---------- End of error output of Backint for SAP DB
(E:\sapdb\BOO\db\backups\logs\backint.err)----------
...
Not removing Backint for SAP DB's temporary error output file
(E:\sapdb\BOO\db\backups\logs\backint.err).

As those files can not be read in that moment by the DBM Server, it
leaves them in the file system for further analysis by the user.
Unfortunately the user has to be remove the before the next backup
attempt manually.

I guess we need to look into the inheritance of file handles between DBM
Server, Backint for MaxDB and child processes of Backint for MaxDB once
more, as it seems that a killed Backint for MaxDB can still keep a file
inaccessible.

Thank you for your help.

Best regards,
Tilo Heinrich
SAP Labs Berlin

 

 

________________________________

From: Jared Still [mailto:[EMAIL PROTECTED] 
Sent: Thursday, December 22, 2005 7:55 PM
To: Heinrich, Tilo
Cc: [email protected]
Subject: Re: MaxDB backups



I've attached dbm.ebl

Jared



On 12/22/05, Heinrich, Tilo <[EMAIL PROTECTED]> wrote: 

        Hallo Jared,
        
        >When the MaxDB backup fails, it does not cleanup its log files.
        
        Do you mind sending a dbm.ebl to the mailing list, so that I
might be
        able to find out the reason for this?
        
        Thank you,
        Tilo Heinrich
        SAP Labs Berlin
        
        ________________________________
        
        From: Jared Still [mailto: [EMAIL PROTECTED]
        Sent: Wednesday, December 21, 2005 9:10 PM
        To: Heinrich, Tilo
        Cc: [email protected]
        Subject: Re: MaxDB backups 
        
        
        Hi Tilo,
        
        Comments inline:
        
        
        
        
                After the backup got canceled, there should be more
information
        within dbm.ebp, then just the "Waiting... is running"-blocks you
        presented. 
        
                Could you please explain further on the usefulness of
dbm.ebp
        and the number of backups attempts?
        
        
        NetBackup makes 3 attempts to run a job. If the 1st one fails,
it tries
        again. If that one also fails, it will try a 3rd time. 
        After the 3rd failure, NetBackup gives up and returns an error.
        
        When the MaxDB backup fails, it does not cleanup its log files.
        When the next attempt to run the job sees that those log files
still
        exists, it throws an error. 
        Since this second attempt to run has overwritten dbm.ebp, the
log of the
        initial failed backup job is lost.
        
        Here is an example from dbm.epl:
        
        Error thrown by 2nd backup attempt:
        
            Have encountered error -24927: 
                The file 'E:\sapdb\BOO\db\backups\io\backint.output'
already
        exists.
        
        The real error appeared in the previous backup attempt as
recorded in
        dbm.epl.
        
        The number of retries is probably configurable.  I haven't
checked yet, 
        but it would be useful to
        set this to 1 for troubleshooting purposes.
        
        
        
        
                Within dbm.prt only the issued DBM Server commands are
        accumulated not any logs. Did you mean dbm.ebl?
        
        
        Yes, I did mean dbm.ebl.
        
        
        
        --
        Jared Still
        Certifiable Oracle DBA and Part Time Perl Evangelist
        
        
        
        
        




-- 
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist 

Reply via email to