RE: Recover from lost log volume

"David Lane" <[email protected]>
Newsgroups gmane.comp.db.maxdb
Message-ID <FA1A4C212C136B4883E784FCE544BA996E62BE@MERCURY.auriga.bhead.co.uk>
Thanks, Tilo.

Two questions: 
- How does autorecover relate to db_activate? Where the log volume is
lost, do I do:

db_activate recover ...
autorecover 

- Is autorecover available in MaxDB 7.5?

Thanks

David Lane


> -----Original Message-----
> From: Heinrich, Tilo [mailto:[email protected]] 
> Sent: 30 January 2007 12:42
> To: [email protected]
> Subject: RE: Recover from lost log volume
> 
> Hello David,
> 
> As you are using an external backup tool, you might want to 
> give the DBM command autorecover a try.
> 
> Recover_replace must be in the same session as the initial 
> recover_start. So db_activate and "recover_start log..." must 
> not be in the same session. You can even restore the log 
> backups in more then one sesion, but you have the start all 
> these sessions with a "recover_start log...". If you do so, 
> you must be prepared to restore a small parts of the log more 
> then once. The DBM command  db_restartinfo helps you to 
> determine the next log backup needed.
> 
> "save_skipped" means that you tried a log backup that is not 
> needed at the moment.
> 
> Best regards,
> Tilo Heinrich
> SAP Labs Berlin
> 
> -----Original Message-----
> From: David Lane
> Sent: Dienstag, 30. Januar 2007 13:32
> To: Brunzema, Martin; [email protected]
> Subject: RE: Recover from lost log volume
> 
> I'm following a previous thread (I/O error trying to recover 
> log backup, earlier in January), where the sequence that worked is:
> 
> db_connect
> db_activate recover medium data ...
> recover_start logmedium log ...
> 
> Does the whole recovery have to be done in one dbmcli 
> session, or can I do the db_activate in one session, then the 
> recover_start in a second session, then recover_replace in a 
> third session, and so on:
> 
> db_connect
> db_activate recover medium data externalbackupid "EBID" autoignore ...
> db_connect
> recover_start logmedium log externalbackupid "EBID"
> 
> What I see is that the db_activate works, but the 
> recover_start restore of the log hangs. In the database 
> messages (verbose) I see:
> 
>   2007-01-30 12:13:17     Log            0:  no update of LogInfoPage:
> DeviceState = Cleared, LogIsEmpty = true
>   2007-01-30 12:13:17     KernelComm     6:  k38headmaster, Errorcode
> 4306 "save_skipped"
>   2007-01-30 12:13:17     KernelComm     6:  RestoreLog, 
> Errorcode 4306
> "save_skipped"
> 
> I know that the dbmgui can do all this for me, but we have a 
> customer who wants to do restores through our backup console, 
> which uses dbmcli, so I need to work out the right sequence 
> of dbmcli commands.
> 
> Thanks
> 
> David Lane
> 
> > -----Original Message-----
> > From: Brunzema, Martin
> > Sent: 30 January 2007 12:07
> > To: [email protected]
> > Cc: David Lane
> > Subject: RE: Recover from lost log volume
> > 
> > Hi David,
> > 
> > in order to recreate the logvolumes it is not sufficient to restore 
> > data/restore log. Instead of a restore data you has to "reactivate" 
> > the instance with the dbmcli-command
> > 	db_activate RECOVER <your databackupmedium>
> > 
> > The dbmgui offeres this option by using "recovery with 
> initialization"
> > 
> > But be aware that all data in the data- and logvolumes is 
> cleared by 
> > this step.
> > 
> > Kind regards, Martin
> > 
> > > -----Original Message-----
> > > From: David Lane
> > > Sent: Tuesday, January 30, 2007 11:03 AM
> > > To: Hahn, Uwe; [email protected]
> > > Subject: RE: Recover from lost log volume
> > > 
> > > Uwe,
> > > 
> > > Thanks for your reply. I have tried that. 
> > > - Restore full - works ok
> > > - Restore pages - works ok
> > > - Restore log - fails.
> > > 
> > > Dbmcli reports
> > > -24920,ERR_BACKUPOP: backup operation was unsuccessful 
> The database 
> > > was unable to fulfill a request (-902, I/O error).
> > > 
> > > Database messages (verbose) shows:
> > >   2007-01-30 09:51:47     IO         11597:  Open
> > > '/var/opt/sdb/data/MAXDB1/log/DISKL0001' successfull, fd: 17
> > > Thread   0x1A
> > >   2007-01-30 09:51:47 ERR IO/READ    11987:  read error: rc=0,
> > > wanted=8192, 'NO ERROR(0)', try again Then
> > > Thread   0x14 Task  97
> > >   2007-01-30 09:51:49     vattach    11000:
> > > '/var/opt/sdb/data/MAXDB1/log/DISKL0001' T97 failed
> > > Thread   0x14 Task  97
> > >   2007-01-30 09:51:49 ERR IOMan         26:  Attach error on 
> > > Log volume
> > > 1: Could not open volume
> > > Thread   0x14 Task  97
> > > 
> > > The problem is that the log volume
> > > /var/opt/sdb/data/MAXDB1/log/DISKL0001 was lost.
> > > I have tried the log restore with
> > > /var/opt/sdb/data/MAXDB1/log/DISKL0001
> > > missing, and it fails.
> > > I also tried creating an empty
> > /var/opt/sdb/data/MAXDB1/log/DISKL0001
> > > file, but it still fails.
> > > 
> > > I cannot find any way to recreate
> > > /var/opt/sdb/data/MAXDB1/log/DISKL0001.
> > > 
> > > Your help will be much appreciated.
> > > 
> > > Thanks
> > > 
> > > David Lane
> > > 
> > > > -----Original Message-----
> > > > From: Hahn, Uwe
> > > > Sent: 29 January 2007 17:54
> > > > To: David Lane; [email protected]
> > > > Subject: RE: Recover from lost log volume
> > > > 
> > > > Hi David,
> > > > 
> > > > you have to reinstall your database instance with your last
> > > > data- and log-backups.
> > > > 
> > > > Kind regards
> > > > Uwe
> > > > 
> > > > -----Original Message-----
> > > > From: David Lane
> > > > Sent: Montag, 29. Januar 2007 15:43
> > > > To: [email protected]
> > > > Subject: RE: Recover from lost log volume
> > > > 
> > > >  I forgot to mention: I am running MaxDB 7.6 on solaris 9.
> > > > 
> > > > > -----Original Message-----
> > > > > From: David Lane
> > > > > Sent: 29 January 2007 14:37
> > > > > To: [email protected]
> > > > > Subject: Recover from lost log volume
> > > > > 
> > > > > How do I recover from total loss of my log volume? I
> > can restore
> > > > > database backups ok, but when, after restoring the backups,
> > > > I try to
> > > > > change the database to online mode (from
> > > > > dbmgui) I get
> > > > >  
> > > > > -24988 sql error [db_online -f] -902,I/O error.
> > > > >  
> > > > > In knldiag:
> > > > >  
> > > > > 2007-01-29 14:23:20       0x14     11000 vattach
> > > > > '/var/opt/sdb/data/MAXDB1/log/DISKL0001' T92 failed
> > > > > 2007-01-29 14:23:20       0x14 ERR    26 IOMan    Attach 
> > > > error on Log
> > > > > volume 1: Could not open volume
> > > > > 
> > > > > /var/opt/sdb/data/MAXDB1/log/DISKL0001 is the lost log
> > > > volume. I have
> > > > > tried creating the file (just an empty file) and that 
> makes no 
> > > > > difference. I found something in the mailing archive about 
> > > > > 'util_execute clear log' to reinit the log, but that
> > crashes the
> > > > > database.
> > > > >  
> > > > > What am I missing?
> > > > >  
> > > > > Thanks
> > > > >  
> > > > > David Lane
> > > > >  
> > > > >  
> > > > > 
> > > > > 
> > > > 
> > > 
> > 
> _____________________________________________________________________
> > > > > BridgeHead Software is pleased to confirm this e-mail has
> > > > been scanned
> > > > > for viruses by MessageLabs.
> > > > > 
> > > > 
> > > > 
> > > 
> > 
> _____________________________________________________________________
> > > > BridgeHead Software is pleased to confirm this e-mail has been 
> > > > scanned for viruses by MessageLabs.
> > > > 
> > > > --
> > > > MaxDB Discussion Mailing List
> > > > For list archives: http://lists.mysql.com/maxdb
> > > > 
> > > > 
> > > 
> > > 
> > 
> _____________________________________________________________________
> > > BridgeHead Software is pleased to confirm this e-mail has
> > been scanned
> > > for viruses by MessageLabs.
> > > 
> > > --
> > > MaxDB Discussion Mailing List
> > > For list archives: http://lists.mysql.com/maxdb
> > > 
> > > 
> > 
> 
> _____________________________________________________________________
> BridgeHead Software is pleased to confirm this e-mail has 
> been scanned for viruses by MessageLabs.
> 
> --
> MaxDB Discussion Mailing List
> 
> --
> MaxDB Discussion Mailing List
> For list archives: http://lists.mysql.com/maxdb
> To unsubscribe:    
> http://lists.mysql.com/[email protected]
> 
> 

_____________________________________________________________________
BridgeHead Software is pleased to confirm this e-mail has been scanned for viruses by MessageLabs.

-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/[email protected]
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.