Cancel active MaxDB backup

[Former Member]

How can one cancel an active MaxDB backup that was scheduled via DB13?

MaxDB 7.6.04.16

Linux

ECC 6.0

Thanks!

[lbreddemann]

Hi Barry,

actually there's no button "STOP THE BACKUP NOW!".

Anyhow, there is something you can do (besides stopping the database completely).

With your version of MaxDB, the backup process is managed by a user task.

This user task waits for all backup tasks to be finished and will tell the kernel finally if everything went ok or not.

So while the backup is running, you may tell this user task to stop what it is doing right now (waiting for the backup to finish).

In my test db I started a backup - let's check what going on in there right now:

x_cons test show ac

SERVERDB: TEST

ID   UKT  Win   TASK       APPL Current         Timeout Region     Wait
          tid   type        pid state          priority cnt try    item
T32    7  0x288 User       1304 WaitForJob(s)         0 0               331(s)
T46    4  0x4CC BUPmed          AsynWaitWrite         0 0               1804(s)
T47    4  0x4CC BUPvol          AsynWaitRead          0 0               1804(s)

Ah, yes... we see one User task currently beeing active with "WaitForJob(s)"... what might that be, hmm?

And there are also two other tasks - BUPmed und BUPvol - the siamese twins of the MaxDB backup. BUPvol is the guy reading our precious data from the Dataarea, see the "AsynWaitRead" thing up there...

BUPmed on the other hand is putting all that data to the backup medium. So far everything pretty boring.

Now, let's tell Mr. T32 to stop the hard "waiting for other people's work to be finished" stuff.

x_cons test kill 32

AND BANG!

The backup is gone! Of course we'll find a big fat error message in all those logfiles.

E.g. KNLDIAG shows us:

0x288         1 Savepoin Savepoint (SaveData) started by T32
0x4CC     20004 Pager    SVP(1) Start Write Data
0x4CC     20005 Pager    SVP(1) Stop Data IO, Pages: 0 IO: 0
0x4CC     20006 Pager    SVP(2) Wait for last task: 70
0x4CC     20007 Pager    SVP(2) Stop Wait for last task, Pages: 0 IO: 0
0x4CC     53070 SAVPOINT B20PREPARE_SVP: 79
0x4CC     20008 Pager    SVP(3) Start Write Data
0x288     19633 CONNECT  Connect req. (T34, Node:'', PID:2424)
0x288     19651 CONNECT  Connection released, T34
0x4CC     20009 Pager    SVP(3) Stop Data IO, Pages: 2 IO: 2
0x4CC     20010 Pager    SVP(3) Start Write Converter
0x4CC     20011 Pager    SVP(3) Stop Converter IO, Pages: 16 IO: 16
0x4CC     53071 SAVPOINT B20SVP_COMPLETED: 79
0x288     20095 Log      New DBIdentifier set (winxpvm.chello.at:TEST_20081101_000140)
0x5F0     19625 IO       Async I/O thread started, 'C:\sapdb\data\TEST\data\DISKD0001
0x608     19625 IO       Async I/O thread started, 'c:\funny_backup_file
0x4CC     52101 SAVE     Filetype: file
0x4CC     52024 SAVE     200 pages -> "c:\funny_backup_file"

0x4CC     52024 SAVE     5000 pages -> "c:\funny_backup_file"

0x4CC     52024 SAVE     10400 pages -> "c:\funny_backup_file"

0x4CC     52024 SAVE     12200 pages -> "c:\funny_backup_file"

0x4CC     52024 SAVE     15600 pages -> "c:\funny_backup_file"

0x4CC WRN 52108 SAVE     cancelled by user
0x608     19626 IO       Async I/O thread stopped, 'c:\funny_backup_file'
0x5F0     19626 IO       Async I/O thread stopped, 'C:\sapdb\data\TEST\data\DISKD0001'
0x288 ERR 52012 SAVE     error occured, basis_err 5750
0x288 ERR     3 Backup   Data backup failed
0x288     19634 CONNECT  Connection killed for T32
0x288     19651 CONNECT  Connection released, T32

We see which user task started the savepoint for the backup (a.k.a. SAVEDATA).

We see how the savepoint was completely finished, the backup started and how the pages were written to the backup file.

AND we finally see, that the backup was cancelled by ourself and that the once so useful task 32 had to give his life for the sake of our higher purpose.

Amazing, isn't it?

regards,

Lars

[Former Member]

Hello Lars,

Thank you so much for the amazing answer. It is much appreciated and worked like a charm!!!!

-Barry