Community
 
 
 

CloudPlatform 3.x

343 seguidores
 
Avatar
Administrator

Will not deploy System VMs, spins them up them deletes them

Avatar

Will not deploy System VMs, spins them up them deletes them

Hi,

I'm running Cloudstack 3.0.5 & VMware 5.1 in an advanced network scenario:

6 x Cisco UCS B-200 Blades
Vmware vSphere 5.1
Nexsan E60 Storage Array â 70TB SAS and SATA
Cisco MDS Fibre Switches
Cisco 3750-eâs 10GB expansion ports

VC01 is running in the same network as the Management Server & Primary Storage 10.100.1.0/24
Storage is hosting the secondary storage NFS 10.205.1.0
Public isn't in use but does exist as a network - 10.210.1.0
Guest01 - 10.215.1.0

Management server has 2 NICs (10.100.1.206 and 10.215.1.10)
I have management data centre and a Guest data centre each with 1 host in for now. I have created all the networks on the management data centre host and just left the guest data centre host blank for CloudPlatform to create them itself, is that correct??

I'm getting errors like this in the management server log:

java.lang.Exception: Unable to find vSwitchCE_MGMT

CE_MGMT is a VM port group I manually created in the VM Management host and labelled it in the zone creation wizard.
Currently there is a v-2-VM which is powered off and a s-*-VM which keeps deleting itself and spinning back up with a higher number on the Guest Host. They are currently residing in the cloud.private.untagged.0.1-vSwitch0 along with the ssytem template.

Please could someone have a look and see what's going wrong?

Attached Files


Tom Cotton MEMBERS 15 November 2012 - 17:59 PM
4 comentários
0

Por favor, entrar para comentar.

 
 

Previous 4 comentários

Avatar
Administrator
Avatar

Will not deploy System VMs, spins them up them deletes them

Hi, traffic labels in CloudStack should be configured with vSwitch names, not port groups. Also, vSphere 5.1 is not supported yet so even if you fix the traffic label, there may be other problems.

Best regards,

{color:#555555}Kirk Kosinski{color} !http://www.linkedin.com/favicon.ico!
{color:#999999}MCITP: EA / VA / EDA7, VCP 4 / 5, CCA{color}


Kirk Kosinski CITRIX EMPLOYEES 19 November 2012 - 04:29 AM
Ações de comentário Permalink
Avatar
Administrator
Avatar

Hi Kirk,

Thanks for that, the V Console VM now stays powered on which looks better.Although I'm still getting these errors (full log attached):

2012-11-20 08:13:03,158 WARN [vmware.resource.VmwareResource] (DirectAgent-97:esx03.workspace.local) Unrecognized broadcast type in VmwareResource, type: LinkLocal. Use vlan info from labeling: untagged

The SSvM however doesn't stay up.

Attached Files


Tom Cotton MEMBERS 20 November 2012 - 17:44 PM
Ações de comentário Permalink
Avatar
Administrator
Avatar

There seems to be a problem with the vSwitch configuration. It may not be connected to a physical NIC, or the switch port the NIC is connected to may not have the right configuration (e.g. missing VLANs or incorrect native VLAN). In this error:

> 2012-11-20 08:13:22,775 INFO \[vmware.resource.VmwareResource] (DirectAgent-97:esx03.workspace.local) Could not connect to 10.100.1.225 due to java.net.NoRouteToHostException: No route to host

It appears the management server cannot reach 10.100.1.225, which appears to be a system VM. Configure the vSwitch configuration so the CloudStack private / management network is reachable from the CloudStack management server. If it needs to be on a VLAN, add the VLAN ID to the traffic label (it should be "vswitchname,vlanid", e.g. "vSwitch0,100" for VLAN 100 on vSwitch0).

Best regards,

{color:#555555}Kirk Kosinski{color} !http://www.linkedin.com/favicon.ico!
{color:#999999}MCITP: EA / VA / EDA7, VCP 4 / 5, CCA{color}


Kirk Kosinski CITRIX EMPLOYEES 26 November 2012 - 07:04 AM
Ações de comentário Permalink
Avatar
Administrator
Avatar

Thank you very very much Kirk. I now have a functional deployment!!!!


Tom Cotton MEMBERS 27 November 2012 - 08:17 AM
Ações de comentário Permalink

Top Contributors