cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From CSG - Ashley Lester <ash...@computer-services.com.au>
Subject RE: Host Down - CS Reports VMs Running on failed host
Date Sat, 19 Oct 2013 21:49:21 GMT
The email alerts from CS show that it did try HA 

Unable to restart csgtest which was running on host name: CSG-XS-H2(id:6), availability zone:
1, pod: Pod1

What would be the effect on the running VM's for the following : toolstack restart first (might
want to kill hung processes like xapi). if that doesnt get you far emergency reset pool master
to host 2




-----Original Message-----
From: Ahmad Emneina [mailto:aemneina@gmail.com] 
Sent: Sunday, 20 October 2013 7:32 AM
To: Cloudstack users mailing list
Subject: Re: Host Down - CS Reports VMs Running on failed host

CloudStack initially queries the host (host1 in this case), but if it fails to connect to
that host.. it will do further investigation by querying the vm from another host in the cluster.
Thats probably why it hasnt tried to HA your vm's, it can see they're alive from host2.


On Sat, Oct 19, 2013 at 2:22 PM, Eric Tykwinski <eric-list@truenet.com>wrote:

> Ashley,
>
> I was just playing with this in my own lab...
> I'm currently running XenServer 6.2.  From what I can tell, CloudStack 
> relies on the underlying hypervisor to do HA.
> I'm thinking CloudStack uses the restart.retry.interval variable to 
> test whether the host/vm is alive even if the underlying hypervisor 
> has moved the vm.  Coming from a background in ESXi and used to the 
> single ping delay, I'm a bit confused on optimal settings in a work environment.
>
> l variable,
> Sincerely,
>
> Eric Tykwinski
> TrueNet, Inc.
> P: 610-429-8300
> F: 610-429-3222
>
> On Oct 19, 2013, at 4:48 PM, CSG - Ashley Lester < 
> ashley@computer-services.com.au> wrote:
>
> > Hello,
> >
> > I have a situation where CS has emailed a alert saying host 2 (was 
> > the
> pool master) is down:
> > I have found the following :
> >
> > Checking finds XenServer Host 2 running but not accessible in the 
> > pool,
> its console show "No VMs Running on this host"
> > CouldStack reports Host 2 down but also reports that some  VMs are
> running on this host
> > The VMs that CS reports as running on the host2 are still actually
> running on host2, I can log into them and they are not present on any 
> other host
> >
> > Asking CS to move the VM to another host fails.
> >
> > Can anyone help with some advise ?
> >
> > I was considering a toolstack reboot on the "failed host"
>
>
>

Mime
View raw message