Flakes in `chain_query_interface` tests

[dorianvp]

A recent change made some flakes to appear more frequently. Upon further investigation (see #610) it turns out that, after removing the sleep and implementing a proper polling mechanism, the NFS drains when 100+ blocks are mined, and this causes a bunch of tests to fail if they hit this window.

See this test for a repro of the issue. The fix should make that test pass.

[nachog00]

is this fixed? @dorianvp @zancas

[dorianvp]

is this fixed? @dorianvp @zancas

This issue and the linked PR mostly describe the weird behavior around the NFS. My understanding is that a workaround was implemented in #638... We can either rename this issue to refer to that more explicitly, or mark this as completed and open another one targeting the core issue.

[nachog00]

would you expect #610 to pass if it got #638 merged into it?