Hi, "avoid set" messages indicate some error occurred earlier in the
job, so to find the real error you need to continue looking (i.e. scroll
up in the log). I did this and found:
2014-11-26 16:14:47,834 DEBUG [c.c.a.t.Request]
(AgentManager-Handler-11:null) Seq 33-995164179: Processing: { Ans: ,
MgmtId: 90581290632, via: 33, Ver: v1, Flags: 110,
[{"com.cloud.agent.api.Answer":{"result":false,"details":"ssh: connect
to host 169.254.3.192 port 3922: No route to host","wait":0}}] }
The Virtual Router (r-167-VM) has some kind of problem that is
preventing the host from connecting to it over SSH. You can log into
the VR console to see if you can find why the host cannot reach it on
the NIC with 169.254.3.192, but the fastest solution is probably to just
recreate the VR (assuming there are no other problems in the environment).
There are two ways to recreate a VR.
1) Destroy it in the CloudStack UI or API and start a VM in the network
(such as the affected VM, i-31-168-VM).
2) Restart the network with cleanup enabled, also via the UI or API.
The eventual capacity error is misleading, but it sort of makes sense
when you look at what happened. After the VR problem blocked CloudStack
from starting the VM on the selected host, the pod was put into the
avoid set and CloudStack looked for somewhere else to put the VM. It
could not find capacity elsewhere and thus reports insufficient capacity.
Best regards,
Kirk
On 11/26/2014 08:16 AM, veera wrote:
> Hi,
>
> I have working cloudstack setup with 1 zone - 1 pod - 2 clusters.
> Now I am getting the error "Unable to start a VM due to insufficient
> capacity" when i try to start a VM.
>
> the log says "The specified pod is in avoid set, returning".
> Any hints?
>
> thanks.
