Restarting cloudstack-agent and cloudstack-management made the inactive
management node notice go away but the instance is still stuck in an
"expunging" state.  How can I get rid of it?


On Thu, Feb 20, 2014 at 10:55 AM, Ian Young <[email protected]> wrote:

> I noticed that there are 12 records in the cloud.mshost table, all of
> which have an "Up" state.  I only have one management server.  Should I
> delete the other 11 records?
>
>
> On Thu, Feb 20, 2014 at 10:20 AM, Ian Young <[email protected]>wrote:
>
>> Yesterday I tried to start an existing instance but it failed.  Since it
>> was basically a brand new installation, I just decided to destroy it and
>> start over.  However, it stayed in an "expunging" state and remains so
>> today.  I cannot create new instances now.  The management-server.log shows
>> numerous messages like this:
>>
>> 2014-02-20 10:04:04,442 DEBUG [cloud.cluster.ClusterManagerImpl]
>> (Cluster-Heartbeat-1:null) Detected management node left, id:11,
>> nodeIP:192.168.100.6
>> 2014-02-20 10:04:04,442 INFO  [cloud.cluster.ClusterManagerImpl]
>> (Cluster-Heartbeat-1:null) Trying to connect to 192.168.100.6
>> 2014-02-20 10:04:04,442 INFO  [cloud.cluster.ClusterManagerImpl]
>> (Cluster-Heartbeat-1:null) Management node 11 is detected inactive by
>> timestamp but is pingable
>>
>> The management server and the hypervisor host are the same machine
>> (budgetary constraints necessitated this setup) so, obviously, it should be
>> able to connect to itself.  What is this timestamp it's referring to?  Is
>> it simply a matter of updating this so the management server is no longer
>> considered inactive?
>>
>
>

Reply via email to