What happens if the controller has a network split?

asked 2014-03-24 20:48:15 -0600

andrewklau
1 ●2 ●3 ●3

My question, in this scenario where the compute nodes lose connection to the controller primarily the HEAT API?

Scenario diagram:

    world
      |
  Networker
      |
 (switch2)
      |
  Compute
      |                       
   (switch1)------------- Controller

Background:

I've been playing around with heat and instance HA with the new HOT template formatting. I have my generic template here, https://raw.githubusercontent.com/andrewklau/openstack-heat-templates/master/base-instance-iha.yaml (https://raw.githubusercontent.com/and...) it's not properly working yet (validation issues), but the main concept should make sense.

In the case scenario where above (switch1) were to fail and go offline for 30 minutes. If I understand correctly, most services shouldn't get effected but the heat component (primarily the IHA) would get hit pretty hard. Are there any case scenarios regarding this?

Thanks.

edit retag flag offensive close merge delete

1 answer

Sort by » oldest newest most voted

0

answered 2014-04-13 22:32:34 -0600

flag of New Zealand

584 ●8 ●11

I think what you're saying is that if the server cannot send its heartbeat due to networking issues rather than server issues then the server will be unnecessarily restarted. Is this correct?

If so, I can't see any obvious way of avoid this, but the consequences are not too serious.

edit flag offensive delete link

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2014-03-24 20:48:15 -0600

Seen: 144 times

Last updated: Apr 13 '14

Related questions