cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrija Panic <andrija.pa...@gmail.com>
Subject Acs-management fencing VMs, on agent unavailable
Date Thu, 17 Mar 2016 14:22:05 GMT
Hi guys,

we had some strange (I would say...) happening last night.

We have 2 node keepalived "cluster" with haproxy providing
agent-to-management-server connections.

We did some LB maintance, so we restarted keepalived (VIP moved to another
haproxy node...), so there was I would say 2-3 sec network downtime betwen
management server and agents.

Management server freaked out, and started Fencing bunch of VMs and VRs...

Any explanation on some time limit that when exceded, management will start
stopping VMs ?

WE did this previously with no consequences as far as I remember...

How are we supposed to more "properly" do LB maintance so management server
doesnt go wild ?

Thanks for any input!

-- 

Andrija Panić

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message