Info required about ontape restore
Posted in 2008
Topics: Backup & Restore, Server Administration, Logging & Checkpoints
Hello,
Query #1
I have completed a restoration of compressed ontape level 0 backup on my test
server using the backup and restore filter /bin/gzip.
the compressed level 0(on file system) used for restoration was residing on
remote server.
After the level 0 was restored & before prompting for the restoration of level
1
========================================================
Continue restore? (y/n)y
Do you want to back up the logs? (y/n)n
Remote tape report:
1174565+1 records in
1174565+1 records out
Restore a level 1 archive (y/n) n
Do you want to restore log tapes? (y/n)n
/opt/informix/ids-11.10.FC2W1/bin/onmode -sy
Program over.
========================================================
Pls let me know what does the Remote tape report: means
ie
1174565+1 records in
1174565+1 records out
what exactly does it means.
#############################################################################
Query #2
when the restoration is in progress there are no dirty buffers (correct me if
i am wrong ) in o/p of onstat -R.
but once the log restoration is cancelled then
15:02:13 No logical log restore will be performed.
15:02:13 Clearing the physical and logical logs has started
15:05:25 Cleared 13366 MB of the physical and logical logs in 191 seconds
15:05:25 Physical Recovery Started at Page (4:759119).
15:05:25 Physical Recovery Complete: 0 Pages Examined, 0 Pages Restored.
15:05:25 Logical Recovery Started.
15:05:25 10 recovery worker threads will be started.
15:05:27 Logical Recovery has reached the transaction cleanup phase.
15:05:27 Logical Recovery Complete.
0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks
and now when i check onstat -R , i can see many Dirty buffers.what is
happening, when
"Physical Recovery Complete: 0 Pages Examined, 0 Pages Restored"
And "Logical Recovery Complete.
0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks".
Pls explain me what makes the buffers Dirty once the restoration of level 0 is
completed & no level 1 or log restoration is carried out .
Thanking you all in advance.
VIKAS HIVARKAR wrote:
That's the output from dd which was used to pipe the compressed file to
gzip.
> Hello,
>
> Query #1
> I have completed a restoration of compressed ontape level 0 backup on my test
> server using the backup and restore filter /bin/gzip.
>
> the compressed level 0(on file system) used for restoration was residing on
> remote server.
> After the level 0 was restored & before prompting for the restoration of
level
> 1
> ========================================================
> Continue restore? (y/n)y
> Do you want to back up the logs? (y/n)n
>
> Remote tape report:
>
> 1174565+1 records in
> 1174565+1 records out
>
> Restore a level 1 archive (y/n) n
> Do you want to restore log tapes? (y/n)n
> /opt/informix/ids-11.10.FC2W1/bin/onmode -sy
>
> Program over.
>
> ========================================================
> Pls let me know what does the Remote tape report: means
> ie
> 1174565+1 records in
> 1174565+1 records out
> what exactly does it means.
> #############################################################################
> Query #2
> when the restoration is in progress there are no dirty buffers (correct me if
> i am wrong ) in o/p of onstat -R.
> but once the log restoration is cancelled then
>
> 15:02:13 No logical log restore will be performed.
> 15:02:13 Clearing the physical and logical logs has started
> 15:05:25 Cleared 13366 MB of the physical and logical logs in 191 seconds
> 15:05:25 Physical Recovery Started at Page (4:759119).
> 15:05:25 Physical Recovery Complete: 0 Pages Examined, 0 Pages Restored.
> 15:05:25 Logical Recovery Started.
> 15:05:25 10 recovery worker threads will be started.
> 15:05:27 Logical Recovery has reached the transaction cleanup phase.
> 15:05:27 Logical Recovery Complete.
>
> 0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks
>
> and now when i check onstat -R , i can see many Dirty buffers.what is
> happening, when
> "Physical Recovery Complete: 0 Pages Examined, 0 Pages Restored"
> And "Logical Recovery Complete.
>
> 0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks".
>
> Pls explain me what makes the buffers Dirty once the restoration of level 0
is
> completed & no level 1 or log restoration is carried out .
>
> Thanking you all in advance.
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
>
--
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)
Query #1 has been answered, so I'll take a stab at your query #2.
Ontape level 0 backs up all data pages as they are when the backup is
initiated. These pages include uncommitted data from transactions open at that
point in time. The ontape level 0 also includes the Logical Log entries for
these open transactions. If you choose to not restore logical log records,
Informix will rollback those open transactions as part of the restore.
Providing a clean view of data that was committed at the point in time the
level 0 was initiated and leaving some dirty buffers.
DAVE GRIFFEN wrote:
> Query #1 has been answered, so I'll take a stab at your query #2.
>
> Ontape level 0 backs up all data pages as they are when the backup is
> initiated. These pages include uncommitted data from transactions open at
that
> point in time. The ontape level 0 also includes the Logical Log entries for
> these open transactions. If you choose to not restore logical log records,
> Informix will rollback those open transactions as part of the restore.
> Providing a clean view of data that was committed at the point in time the
> level 0 was initiated and leaving some dirty buffers.
>
Correct enough, almost complete. The archive includes all data pages in
the instance, it also includes the PHYSICAL log pre-images of any pages
modified since the beginning of the archive. At restore time these
pre-images are slapped down over the data pages after the dbspaces are
restored. This insures that the final view of the pages on disk is
exactly as they were at the moment the checkpoint at the beginning of
the archive began. And, yes, the dbspace pages also include the logical
logs as of the beginning of the archives. As you stated, if you do not
choose to roll forward the logical logs archived after the physical
dbspace archive was taken then the physical data restore is complete at
that point, otherwise it completes after the logs have been restored.
After that physical data (and possibly the log data) is restored the
engine is started and normal fast recovery rolls back transactions that
were incomplete as of the start of the archive just as if the engine had
crashed.
--
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)