>> Its urgent
Than you should not (only) rely on a 'best effort' Forum question but reach out to a (Qlik) Support or Consulting organization. Of course you can also try a forum, you never know - but you'd better add some concrete information because us strangers know nothing about your specific issues (Qlik might - from pre-sales talks.)
>> currently getting huge latency for only Oracle Tsys prime EC2 Server about 57 hrs
Did it ever work properly? What changed?
>> only for this source
Well, are the other sources Oracle? Are they roughly the same databases? are they in the cloud as well or local? Do they have similar tables counts, change rates, Change volumes? Container databases (a single Redo being shared for all guest containers (PDB's))
>> Can someone also share Best practices and solutions for large latency.
Source or target latency? I know you strongly suggest this is related to a particular source
Anyway - start with getting some reptask logs with PERFORMANCE LOGGING LEVEL set to TRACE. This will give some details for latency every 30 seconds. And for Oracle sources it will give REDO (ARCH) log read performance indicators. Now for an Oracle Source the default Replicate behaviour is to read the entire Redo (Arch ) log an select relevant changes from there. How much redo is being generate (GB/Hour)? How fast is the EC2 connection (MB/sec? Latency?) .
For slower networks and when selecting a relatively small ( < 20%) part of the change log on might want to consider using the deprecated Logminer which will only transfer changes of actual interest, and not just everything.
Good luck!
Hein
