Hey all, sorry if the question was already answered in another topic, but I couldn’t find any information in the docs if reloading a Nomad agent would cause any Job downtime.
I need to change some configurations in our Nomad cluster (client and server) and I want to be sure that doing a reload wouldn’t affect any running Job allocation.
Any advice is also greatly appreciated.
Hi @rmoretto, the answer depends on how your cluster is configured. Generally speaking restarting a Nomad client or server should not impact already running tasks. There are however several configuration option that can change the behavior of what happens, and when.
stop_after_client_disconnect - configures what happens when a Client loses connection with the servers
max_client_disconnect - also configures what happens when a Client loses connection with servers, but with different behavior
heartbeat_grace, min_heartbeat_ttl, failover_heartbeat_ttl - all impact how long the client agent has to become alive again during a restart
You may also find the upgrade process guide valuable