Problem
At the moment, as I tested, if the master (panel) goes down, the node (worker) gets down as well. This is undesirable in distributed systems. the right patter is keep the worker up and running even when the master is not up
Proposed solution
node execution should not depend on panel availability.
when the panel is up, it accepts new requests and commands,
when the panel is not up, no new request or command is accepted
Alternatives considered
at the moment we can dump the config (xray api ...) connecting the api server
Area
Nodes
Impact
Important
I also saw this env
## Experimental features.
# STOP_NODES_ON_SHUTDOWN = True
please let me know, if this env can do what mentioned above
Problem
At the moment, as I tested, if the master (panel) goes down, the node (worker) gets down as well. This is undesirable in distributed systems. the right patter is keep the worker up and running even when the master is not up
Proposed solution
node execution should not depend on panel availability.
when the panel is up, it accepts new requests and commands,
when the panel is not up, no new request or command is accepted
Alternatives considered
at the moment we can dump the config (xray api ...) connecting the api server
Area
Nodes
Impact
Important
I also saw this env
please let me know, if this env can do what mentioned above