<div dir="ltr"><div>In my openstack cluster (Havana), occasionally, i have some instances can not be ping if reboot many instances simultaneuously.<br></div><div>if i "nova console-log ...", i see a line:<br>"<b>cloud-init-nonet waiting 120 seconds for a network device.</b>"<br>
</div><div>To fix this, i must run "<b>service neutron-plugin-openvswitch-agent restart</b>" on compute-nodes that these instances is hosted.<br></div><div>After restart neutron-agent, i can ping these instances normally. <br>
<br><br></div><div>Therefore, I build another openstack-cluster (Havana) to test, but the problem persist (i test with both VLAN and GRE, both soft_reboot and hard_reboot - but this problem is the same!)<br># for server_id in `nova list | grep ACTIVE | awk '{print $2}'`;do echo $server_id ; nova reboot --hard $server_id; done<br>
or<br># for server_id in `nova list | grep ACTIVE | awk '{print $2}'`;do echo $server_id ; nova reboot $server_id; done<br><br>Check console-log:<br># for server_id in `nova list | grep ACTIVE | awk '{print $2}'`;do echo $server_id ; nova console-log $server_id | grep "waiting 120 seconds"; done <br>
cea03292-5c76-45dd-af35-b9d80b82ac5d<br>37b02ca8-98fc-40c7-a700-f7942d134b65<br>a92051b1-3a1d-4f68-ba77-4d093e916ce4<br><b>cloud-init-nonet waiting 120 seconds for a network device.</b><br>def8f657-572a-4431-902a-c983d0aad7a9<br>
<b>cloud-init-nonet waiting 120 seconds for a network device.</b><br><br></div><div>And I can ping 2 instances: a92051b1-3a1d-4f68-ba77-4d093e916ce4 & def8f657-572a-4431-902a-c983d0aad7a9<br></div><div><br></div><div>
This bug exists quite along time in neutron-agent, but still not be fixed.<br><br></div><div>Can you suggest me a quick fix for that? <br>Thank you very much!<br></div><div><br></div><div>Best Regards,<br></div></div>