aguestuser requested to merge hf-cleanup-sharding into main Oct 22, 2020

context

in !417 (merged) we shipped sharding
in !429 (merged) we tried to cleanup some booboos
but we didn't catch them all! this MR finishes the job. :)

changes

in deploy.yml playbook: pull signald container by referencing signald_0 service (since signald service doesn't exist anymore)
in socket.run() don't invoke pool.stop() when defining it!!! (that causes the socket connection to close as soon as it is opened and subscribe calls to throw errors about trying to write to a stream that has already been closed! why did this not fail in dev? great question!!!!)