Help please...

[waffle]

i don't think it's very likely that your network is the bottleneck in this process. how are they currently linked? to test, you could measure the amount of network activity on the two computers and see if it's pegging out at the max speed of the link. but i'd be pretty suprised if this were the case.

i bet, as you did, it's poor sql performance on the mining computer that's slowing it down. it's either a lack of resources or like you said, some sort of locking problem caused while also importing.

what happens if you tell PA to use the remote, mined database, but you aren't mining at the same time? is it still slow?

if this is the case, then the queries are just running slowly. you will have to throw more hardware at the mining machine. maybe more memory, faster disk, dedicated disk, etc.

if you haven't already i would edit the postgresql.conf on the mining computer and make sure the database is using all the physical memory you can spare for it. the default settings don't cause the db to use a lot of memory. increase the shared_buffers and work_mem settings, do a little research to see how much depending on how much free mem you typically have. this should cause it to hit the disk a little less.