Maybe you should look into this problem first. I am not sure why would you use so many nodes, instead of relying more on harvesters that require far less resources and have way less issues when running.
2 Likes
Those log lines are INFO level, so rather irrelevant. Basically, the farmer/wallet cannot talk to the full node (a secondary problem). You need to scan your logs for ERROR level lines especially (or rather only) those that are coming from the node. The fact that your UI has errors all over it implies that ERROR level log lines coming from the node are already in the log.
Also, I would think that you should see those ERROR lines within minutes in your logs, so really no need to wait an hour. The only delay you can have in the logs is on the harvester level, as it may take some time to scan all drives (with massive harvesters). Node is rather fast, as it primarily deals with peers, and those timeouts are rather short (a minute or two would be my guess).