Hi,
I was trying my CoreOS's PXE setup for bare metal devices and found 
something strange: sometimes the bare metal device can be configured with 
the IP specified in cloud-config-url and public key ingested, but sometimes 
it just cannot.

Looking into journalctl logs, I found that in the failed cases, 
user-cloudinit-proc-cmdline.service exited with failure after 15 retries 
and within every retry, it is complaining for
'network is unreachable'. We have a slow DHCP server and since the DHCP 
packages are broadcasted across several switches using VLAN, it's expected 
that there may be some latency for DHCP to finish (possible more than 
1min...). So I'm asking if I can change the number of the maximum retry 
times? Or can I change the maximum sleep time to be larger than the maximum 
5s?

Thanks,
Philip

Reply via email to