Have 72 trail files of 2 GB each.
With a single replicat am able to only get to 10,000 document inserts per second and a transfer speed of 3 MB/s.
There is no crunch of resource on gg server or on my mongo server.
I tried running 10 parallel replicats which gets me about 80,000 document inserts per second and a transfer speed of 18 MB/s.
This is not sufficient. When i try upload a normal file from source to destination am able to get 130 MB/s. Network is not an issue.
replicat does not even get to 25% of total JVM memory allocated.
What are some parameters i can play with or where could the bottleneck be. Kindly assist
Oralce version 19c
This is my replicat configuration:
REPLICAT MY_REPLICAT
TARGETDB LIBFILE libggjava.so SET property=/data/mongo.props
REPORTCOUNT EVERY 1 MINUTES, RATE
GROUPTRANSOPS 40000
MAP SCHEMA.TABLENAME,TARGET "schema"."collection_08102024"
This my mongo.props:
gg.handler.mongodb.WriteConcern={"w":"1", "wtimeout": "0"}
gg.handler.mongodb.BulkWrite=true
gg.handler.mongodb.upsert=false
jvm.bootoptions=-Xmx8048m -Xms8048m
javawriter.bootoptions=-Xms4048m -Xms4048m
Target is a self-managed 4 shard mongo cluster