Re: RMAN catalog operations taking 128 seconds per file - where is the bottleneck (and no guessing)

On Wed, Oct 13, 2010 at 12:59 PM, Charles Schultz <sacrophyte@xxxxxxxxx>wrote:

> Thanks, Jared, looks like it could quite possibly be unpublished bug
> 7444008<https://supporthtml.oracle.com/ep/faces/secure/km/BugSearch.jspx?bugId=7444008>.
> I hate bugs. Still, I am curious, what exactly is the root problem?
>
>
You might try tracing RMAN.

There are many ways to do this.

From my notes:

*COMMAND LEVEL TRACE*

Trace and debug from levels 1-6, possibly higher.



Experiment to see differences.



configure channel

device type 'sbt'

parms="BLKSIZE=1048576,SBT_LIBRARY=oracle.disksbt,ENV=(BACKUP_DIR=/mnt/rsysdevdb/bkup_2/ts03)"

maxopenfiles 4

trace 6 debug 6;




allocate channel

device type 'sbt'

parms="BLKSIZE=1048576,SBT_LIBRARY=oracle.disksbt,ENV=(BACKUP_DIR=/mnt/rsysdevdb/bkup_2/ts03)"

maxopenfiles 4

trace 6 debug 6;


*COMMAND LINE TRACE*

$OH/bin/rman … debug trace "trace file name"


*DEBUG (lots of output)*

Include slaxdebug after the target



eg. rman target nbu/password@orcl slaxdebug catalog rman/password@rcat




*SYSTEM LEVEL TRACE* (I have not tried this one, do not try on production)


From Tuomas Pystynen



http://www.miracleltd.com/msdbf2007/tp.pdf



RMAN tracing



alter system set service_names = 'TEST, TEST_RMAN' scope=both;



exec dbms_monitor.serv_mod_act_trace_enable(
service_name=>'TEST_RMAN',module_name=> 'backup full datafile',waits=>true);



exec
dbms_monitor.serv_mod_act_trace_enable(service_name=>'TEST_RMAN',module_name=>
'backup incr datafile',waits=>true);



$ rman target=sys/password@test_rman

backup device type 'sbt_tape' tablespace 'TEST';


Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist
Oracle Blog: http://jkstill.blogspot.com
Home Page: http://jaredstill.com

Other related posts: