Ontape restore questions? certainly I am missing some thing!
Posted in 2000
Topics: Backup & Restore, Storage & Space Management, Logging & Checkpoints
I performed a level 0 backup, using ontape, with the following
parameters in the oncofig file.
TAPEDEV /dev/rmt/0 #
TAPEBLK 16 #
TAPESIZE 22000000 ## Log Archive Tape Device
LTAPEDEV /dev/null
LTAPEBLK 16 #
LTAPESIZE 1024000 #
As you notice we do not archive the Log (/dev/null for LTAPEDEV )
since all our inserts and updates to the database are batch loads type.
I tried to restore one chunk from the tape using 'ontape -r - D
my_chunk'.
First I answered 'y' to the question: 'Do you want to restore log
tapes? (y/n) y'. but the restore failed because of invalid /dev/null;
then I answered 'n' to the same question and the restore terminated
with no error; but when I tried to access one database that resides in
that chunk I realize that my chunk becomes inconsistent and was marked
with 'PI-' flag when I run onstat -d.
My questions are:
1 - Why do I need to restore the log tapes since the LTAPEDEV is
directed to /dev/null; I just need a physical backup of the chunk I
archived?
2 - A level 0 archive is not suppose to back up the log?
I changed LTAPEDEV to /dev/rmt/0 and I loaded the same tape and I
tried to restore the chunk but it didn't work. The system kept asking
me to load an other tape over and over, it is like it did not find the
log to be restored and end with the following message "Unexpected end
of log tape (errno 0), continuing... ";
3 - Finally how I can get out from this mess without calling Informix
support? can I perform, now after 2 weeks, a log backup and use it
to restore the chunk.
Thanks a lot for any clarification.
Sent via Deja.com http://www.deja.com/
Before you buy.
With ontape and OnBar, you can always restore the WHOLE system without
needing the logical logs. Then the system is restored to appear as it was
at the start of the backup.
When you wish to restore only one dbspace, Informix MUST process the logical
logs to bring the dbspace up to current values. You cannot restore a single
dbspace to a previous (in-time) position. This is done to prevent data
inconsistencies.
So, since you are not saving your logical logs, your only option is to
restore the entire instance to 'start of backup'.
Maybe you should consider saving the logical logs????
HTH,
Doug
<dankami@my-deja.com> wrote in message news:8nribf$jen$1@nnrp1.deja.com...
> I performed a level 0 backup, using ontape, with the following
> parameters in the oncofig file.
>
> TAPEDEV /dev/rmt/0 #
> TAPEBLK 16 #
> TAPESIZE 22000000 #> # Log Archive Tape Device
> LTAPEDEV /dev/null
> LTAPEBLK 16 #
> LTAPESIZE 1024000 #>
>
> As you notice we do not archive the Log (/dev/null for LTAPEDEV )
> since all our inserts and updates to the database are batch loads type.
>
> I tried to restore one chunk from the tape using 'ontape -r - D
> my_chunk'.
> First I answered 'y' to the question: 'Do you want to restore log
> tapes? (y/n) y'. but the restore failed because of invalid /dev/null;
> then I answered 'n' to the same question and the restore terminated
> with no error; but when I tried to access one database that resides in
> that chunk I realize that my chunk becomes inconsistent and was marked
> with 'PI-' flag when I run onstat -d.
>
> My questions are:
>
> 1 - Why do I need to restore the log tapes since the LTAPEDEV is
> directed to /dev/null; I just need a physical backup of the chunk I
> archived?
>
> 2 - A level 0 archive is not suppose to back up the log?
> I changed LTAPEDEV to /dev/rmt/0 and I loaded the same tape and I
> tried to restore the chunk but it didn't work. The system kept asking
> me to load an other tape over and over, it is like it did not find the
> log to be restored and end with the following message "Unexpected end
> of log tape (errno 0), continuing... ";
>
> 3 - Finally how I can get out from this mess without calling Informix
> support? can I perform, now after 2 weeks, a log backup and use it
> to restore the chunk.
>
>
> Thanks a lot for any clarification.
>
>
>
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
You can call Informix support and ask for the 'Down Systems Department' < or
something like that>. If the Informix engineer can log into your system via
a
modem he/she can change the db/sb 'state' flags. The idea is to 'markup
the chunk' then the engine can 'accept' the dbspace as a viable resource.
Unfortunately, customers are not allowed to use the utility to get them out
of
the bind. It 'tricks' the engine.
Lee
<dankami@my-deja.com> wrote in message news:8nribf$jen$1@nnrp1.deja.com...
> I performed a level 0 backup, using ontape, with the following
> parameters in the oncofig file.
>
> TAPEDEV /dev/rmt/0 #
> TAPEBLK 16 #
> TAPESIZE 22000000 #> # Log Archive Tape Device
> LTAPEDEV /dev/null
> LTAPEBLK 16 #
> LTAPESIZE 1024000 #>
>
> As you notice we do not archive the Log (/dev/null for LTAPEDEV )
> since all our inserts and updates to the database are batch loads type.
>
> I tried to restore one chunk from the tape using 'ontape -r - D
> my_chunk'.
> First I answered 'y' to the question: 'Do you want to restore log
> tapes? (y/n) y'. but the restore failed because of invalid /dev/null;
> then I answered 'n' to the same question and the restore terminated
> with no error; but when I tried to access one database that resides in
> that chunk I realize that my chunk becomes inconsistent and was marked
> with 'PI-' flag when I run onstat -d.
>
> My questions are:
>
> 1 - Why do I need to restore the log tapes since the LTAPEDEV is
> directed to /dev/null; I just need a physical backup of the chunk I
> archived?
>
> 2 - A level 0 archive is not suppose to back up the log?
> I changed LTAPEDEV to /dev/rmt/0 and I loaded the same tape and I
> tried to restore the chunk but it didn't work. The system kept asking
> me to load an other tape over and over, it is like it did not find the
> log to be restored and end with the following message "Unexpected end
> of log tape (errno 0), continuing... ";
>
> 3 - Finally how I can get out from this mess without calling Informix
> support? can I perform, now after 2 weeks, a log backup and use it
> to restore the chunk.
>
>
> Thanks a lot for any clarification.
>
>
>
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>
Related threads
- IDS 10 table-level restore
- Informix Development Webinar December 11, 2007
- ontape -p/r with changed ROOTPATH
- Migrate from HP PA-RISC to HP ITANIUM by ontape