If a node goes to sleep (or is simply not scheduled?) for some period of time, it becomes impossible to communicate with the node via RPC and maybe also P2P

Description

I set up a few nodes using the cordapp-example, following these instructions: https://docs.corda.net/head/tutorial-cordapp.html

I left the three nodes and springboot server running over night on my laptop (Mac), which went to sleep (ish... think powernap is enabled and so the node was able to run a bit during the night but probably not much).

In the morning, the nodes were still listening on their RPC ports but it was impossible to connect using the Node Explorer.

The nodes themselves were still running, sort of - in that they spotted and logged if I copied a new node-info into the additional-node-infos directory. But they were completely unresponsive to RPC (and nothing was logged by the node when the explorer tried to connect)

It seems the only way to recover from this is to restart the nodes.

The larger attached log file shows the node from startup through laptop sleep and into the evening. The smaller one captures some of my experiments from this morning.

Status

Assignee

Viktor Kolomeyko

Reporter

Richard G Brown

Labels

Feature Team

Enterprise Robustness

Story Points

None

Fix versions

Ported to...

Corda 4.3

Priority

Medium

Affects versions

Configure