VM down time during VM migration

asked 2016-07-27 21:01:50 -0500

fifi gravatar image

updated 2016-10-11 23:18:40 -0500

I'm looking for an accurate way to calculate the VM downtime during live migration.

I know there is a solution using ping -i with very short intervals and then multiply the number of lost ping (ICMP) packets to the interval.

This is an easy and fast solution but is not accurate. Sometimes, the percentage of lost packets is 0 while the number of received ICMP packets is less than icmp transmitted packets. Besides, I observed that that there was no packet loss during pinging a migrating VM while my running services on that VM became irresponsive during live migration, which shows that using ping to calculate downtime is not accurate enough.

I would like to ask if there is a better solution to calculate downtime for a migrating VM in Openstack.

edit retag flag offensive close merge delete