Hello. After 6.42.1 (Current) is installed on the Cloud Hosted Router, the VM starts periodically to be unavailable through the network interfaces, the error is preceded by a message in the dhcp, critical, error :: dhcp-client on WAN lost IP address XX.XX.XX.XX - lease expired.
We’re experiencing a similar issue on Hyper-V, albeit without the DHCP component.
We’re using CHR routers to front customer cloud environments, to route public and private traffic. Periodically these routers are deciding that the private interfaces aren’t working; The symptom being that the next-hops associated with these interfaces are marked as unavailable when they are online. Restarting the CHR router magically resolves the issue. This has only started since the 6.42.1 update, would advise against upgrading in your on Hyper-V.
A workaround is to downgrade to 6.40.8 or add a dhcp renew script (dhcp unicast fails, force renewal will fallback to dhcp broadcast sooner - before release):
/system scheduler
add comment="work around for dhcp-client on [interface] lost IP address x.x.x.x - lease expired" \
interval=25m name=dhcp-renew on-event="/ip dhcp-client renew [ find ]" policy=read,write \
start-time=startup
Tune the interval to renew as late as possible but a few minutes before regular dhcp lease timeout.