Re: RMAN problems.

Alan

Line 1 and 2 are both actually "line 1", so don't need a semi-colon in the middle of them. Line 2 specifies Legato-specific parameters to the allocate channel command (telling it which NSR Server to use). Interesting that the channel type is 'disk' though....

Line 4 has a typo on it (recovert should be recover). "Recover" basically performs any rollforward operation so yes - you would need a RESTORE TABLESPACE TB1; line prior to issuing it if the datafile is totally lost.

RELEASE CHANNEL is good coding practice. RMAN should forceably release channels when you end the run block if you don't do it yourself.

Rate of required catalog resyncs is dependant upon a number of things (but mainly log switch frequency) - you should always resync whenever a physical change is made and at least as frequently as your controlfiles are set to cycle (CONTROL_FILE_RECORD_KEEP_TIME init.ora setting)

Hope this helps

Brian

"Alan" <alan_psb_at_yahoo.com> wrote in message news:38E186DE.D9475AC6_at_yahoo.com...
> Hello all,
>
> I got the following script (for RMAN) from the former DBA in my
> company. I don't know anything (or even try) about RMAN and my manager
> want me to test the script. After reading the script, I found that the
> script MAY have problems compare with the book examples (Oracle8 Backup
> & Recovery Handbook):
>
> The script is used to recover the loss of data files (recover the
> tablespace):
>
> run {
> allocate channel dev1 type disk' (1)
> parms 'ENV=(NSR_SERVER=backup)'; (2)
> sql 'alter tablespace tb1 offline immediate'; (3)
> recovert tablespace tb1; (4)
> sql 'alter tablespace users online'; (5)
> }
>
> 1. Should all the statements in the scripts end with ';'? (in statement
> 1, no ;)
> 2. What is the meaning of line (2)?
> 3. Iit should have one more line before line (4), i.e. restore
> tablespace tb1, am I right?
> 4. All the examples I found in the books are end with 'release
> channel..'. What happen if no such statement in the script?
>
> One more question, could anyone give me hints of determine the
> intervals of synchronization with the target database?
>
> Thanks,
> Alan
>
Received on Wed Mar 29 2000 - 09:54:33 CST