This content has been marked as final. Show 3 replies
It is not saying there is an error, it is just saying your apply site is slow.
Tune the apply site, look for full table scan or any other reason why your machine is slow.
If you don't know what to do, browse this forum.
The question arise quite often and I posted guidelines and queries to sort it out.
You can tune your apply server by tracing on the apply process. To do this: identify the PID (Process ID)
select SID from V$STREAMS_APPLY_SERVER;
With above SIDs, query V$PROCESS to get the PIDs:
select a.pid, b.sid
from v$process a, v$session b
where a.addr=b.paddr and b.sid in(the list of sids from above);
With the PIDs found:
$ sqlplus /nolog
connect / as sysdba
oradebug setorapid <PID>
oradebug Event 10046 trace name context forever, level 12
wait for 10-15 mins
oradebug tracefile_name -- gives pathname / filename to file being written to
oradebug Event 10046 trace name context off
run TKPROF on trace file
you may have to add some indexes on your tables . If possible, each table for which changes are applied by an apply process should have a primary key.
read more in
Section Substitute Key Columns
Edited by: user596799 on Apr 17, 2011 8:24 AM