I have a streams replication set up between two oracle 9.2.0.8 instances as sources and an 11g instance as a destination, processing a small subset of data from each database (let it be one table with no unsupported types and all supplementary logging enabled for this case).
Each source instance has a capture and propagation process. One of the instances works as expected, another one has issues with capture. Here's what I see in v$streams_capture:
17:09:25 STRMADMIN > select total_messages_enqueued, elapsed_capture_time, elapsed_rule_time, elapsed_enqueue_time, elapsed_lcr_time
17:10:04 2 from v$streams_capture;
TOTAL_MESSAGES_ENQUEUED ELAPSED_CAPTURE_TIME ELAPSED_RULE_TIME ELAPSED_ENQUEUE_TIME ELAPSED_LCR_TIME
----------------------- -------------------- ----------------- -------------------- ----------------
0 682521 0 0 337
Elapsed: 00:00:00.16
I infer that the logminer process actually processes a number of redo log entries (elapsed_capture_time=682521), creates LCR objects (elapsed_lcr_time=337), but for some reason never evaluates the rule set (elapsed_rule_time=0), and therefore has no objects to enqueue (elapsed_enqueue_time=0).
So, at this moment I actually ran out of ideas why could that happen. Tried several approaches up to full streams reinstallation without any luck. Probably miss something very close.
Any ideas on this?
Thanks.
select capture_name, STATUS, ERROR_MESSAGE from dba_capture;And any trace or log containing the error message? – Threaten Dec 28 '12 at 17:46