Umm... Isn't a 50 million row export / import usually an hour long task... tops.. I mean assuming you have the scripts?
Dump the first DB, import into the second? Maybe a couple ETL scripts in between to massage the data?
This task sounds like monodb_dump export_db (or whatever they call their backup tool) | etl.sh | psql -d import_db. Of course shell scripts probably aren't 'webscale' enough to handle 150 GB of data.
Protip: When IO is saturated, make the IO sequential before ever even thinking about threads. Exactly how are more CPUs supposed to make the disk faster? The best you'll do is exploit the command queue more fully.
Dump the first DB, import into the second? Maybe a couple ETL scripts in between to massage the data?
This task sounds like monodb_dump export_db (or whatever they call their backup tool) | etl.sh | psql -d import_db. Of course shell scripts probably aren't 'webscale' enough to handle 150 GB of data.
Protip: When IO is saturated, make the IO sequential before ever even thinking about threads. Exactly how are more CPUs supposed to make the disk faster? The best you'll do is exploit the command queue more fully.