Issue running migration against DB with sslmode=require

[lossyrob]

Running migrations against a --dsn that has ?sslmode=require (or with PGSSLMODE=require), I get:

ERROR:asyncio:Fatal error on SSL transport
protocol: <asyncio.sslproto.SSLProtocol object at 0x7f4882e35e20>
transport: <_SelectorSocketTransport closing fd=8>
Traceback (most recent call last):
  File "/opt/conda/lib/python3.8/asyncio/selector_events.py", line 910, in write
    n = self._sock.send(data)
OSError: [Errno 9] Bad file descriptor

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/opt/conda/lib/python3.8/asyncio/sslproto.py", line 685, in _process_write_backlog
    self._transport.write(chunk)
  File "/opt/conda/lib/python3.8/asyncio/selector_events.py", line 916, in write
    self._fatal_error(exc, 'Fatal write error on socket transport')
  File "/opt/conda/lib/python3.8/asyncio/selector_events.py", line 711, in _fatal_error
    self._force_close(exc)
  File "/opt/conda/lib/python3.8/asyncio/selector_events.py", line 723, in _force_close
    self._loop.call_soon(self._call_connection_lost, exc)
  File "/opt/conda/lib/python3.8/asyncio/base_events.py", line 719, in call_soon
    self._check_closed()
  File "/opt/conda/lib/python3.8/asyncio/base_events.py", line 508, in _check_closed
    raise RuntimeError('Event loop is closed')
RuntimeError: Event loop is closed

I'm able to connect to the database using the same dsn using psql. Any ideas?

[lossyrob]

.. It seems that pypgstac did run the migration, but perhaps this is an error that happens on closing?

I'm not seeing any of the logging output when running pypgstac migrate - perhaps that should be output another way?

[bitner]

Looks like this is related to MagicStack/asyncpg#737

[bitner]

I think this might be fixed in asyncpg 2.3
MagicStack/asyncpg#720

I'll try updating the asyncpg version pin.