The run-client.sh script will trigger a reboot of all clients using the bootstrap nodes as the only initial peers.
As this happens every 1800 seconds (30 minutes), it should be expected that nearly every client in the network will be restarting within 30 minutes of the update. For an example network with 10K clients, this would induce a load of ~333 clients per minute on the bootstrap clients.
Proposals that might help reduce network load:
pull only on tags/releases
require manual user input
manual opt-in for reboots on minor changes
increase duration from 30 minutes
This could impact the expectations for client sync test case 3 (900 fringe clients)
💥 Proposal
The run-client.sh script will trigger a reboot of all clients using the bootstrap nodes as the only initial peers.
As this happens every 1800 seconds (30 minutes), it should be expected that nearly every client in the network will be restarting within 30 minutes of the update. For an example network with 10K clients, this would induce a load of ~333 clients per minute on the bootstrap clients.
Proposals that might help reduce network load:
This could impact the expectations for client sync test case 3 (900 fringe clients)