Closed goodboy closed 1 year ago
šš¼ to da CI gawdz
nothing like hanging for 40 mins
THERE. moved all oustanding bullets to new follow up issues šš¼
So whoever decides to review this once more and decides we gud, please merge. If not i'll land later tonight š„
Heh, i can tell y'all aren't testing non-disti mode.. š
Move all actor-service control APIs and the service manager (basically all the stuff that was in
piker._daemon
) into a newpiker.service
sub package; this will serve (get it..) as the high-level actor-service orchestration sub-system, the business logic forpiker
's distributed architecture and runtime deployment.Of note,
.service._ahab
which repair a regression: a silent startup failure withmarketstore
due to the log processing poll sleep period change..ahabd
supervisor taskpikerd
logging correctly to the docker super code in general.._ahab.py
docker supervisor into the sub-package as well since it is literally an external, containerized service supervisor š.data.elastix
and.marketstore
mods here as well until we figure out how to formalize our storage layer subsystems, and since most of that code is related to db interaction/mangement vs. data processing.ToDo:
._daemon.py
module in each sub-system package?.service._daemon
pikerd.data._daemon
contains the routines for spawning thebrokerd
actor?~ => moved to #488Services
api into it's own module?_runtime.py
? ( chose.service._actor_runtime.py
instead )._exec.py
stuff (which is more or less just Qt runtime startup) into this pkg as well?~ => also moved to #488tmp_dir
fixure for our config dir path during testingtractor._state._runtime_vars