cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kirk Kosinski <kirkkosin...@gmail.com>
Subject Re: Cluster With two XenServer
Date Fri, 22 Mar 2013 20:40:56 GMT
Ahmad is right, the most common cause for this is switch
misconfiguration.  Make sure the XS hosts are connected to switch ports
in trunk mode and that all the necessary VLANs are allowed on those
ports.  If the switch ports are definitely configured correctly, confirm
that the XS hosts are connected to the correct ports (and not, e.g. one
in access mode).

If you are using bonding on the XS hosts, make sure that the switch
ports are configured accordingly.  The default SLB mode and the
active-passive modes will probably not work if the switch ports are
configured for LACP, Etherchannel, 802.3ad, etc..  If the NICs in the
bond are connected to different non-stacked switches, use active-passive
and not SLB.

Other possible problems include:
Broken drivers - Check for any driver updates for your XenServer
version.  If that doesn't help or doesn't apply, try a different NIC
(Intel or Broadcom, with updated drivers if applicable).
Open vSwitch problems - Change to bridge back-end temporarily (run
xe-switch-network-backend bridge and reboot both hosts) to see if the
problem is fixed.  If that helps, install all XS patches supported by
CloudStack and try Open vSwitch again.

Best regards,
Kirk


On 03/22/2013 12:10 PM, Ahmad Emneina wrote:
> Hey Tawfq,
> 
> It could be a switch configuration issue. Make sure you trunk what you
> defined for your zone vlans, to each physical switch port the xen hosts are
> connected to. The guests get their private l2 network created by pulling
> from a pool you define in the zone vlan range...
> 
> 
> On Fri, Mar 22, 2013 at 12:06 PM, tawfiq zidi <tawfiq.zidi@gmail.com> wrote:
> 
>> Yes, this deployed with advanced configuration. Even , i checked the setup
>> configuration and i found that guest traffic is enabled.
>>
>> The problem that router DHCP discover available only within the hosts not
>> all the cluster , so the other XenServers in the pool can't see anything
>> from the router.
>>
>> So , is there any others configurations have to be done for XenServers
>> Pool ?
>>
>>
>> On Fri, Mar 22, 2013 at 7:05 PM, Ahmad Emneina <aemneina@gmail.com> wrote:
>>
>>> Is this using a basic or advanced zone. This is definitely NOT how it
>>> should work :)
>>>
>>> VM's across a zone, within the same account, should be able to reach each
>>> other... when deployed to the same network.
>>>
>>>
>>> On Fri, Mar 22, 2013 at 10:53 AM, tawfiq zidi <tawfiq.zidi@gmail.com>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> I've Created XenServer Cluster with two Host ( XS 6.0.2 ) , the system
>>> VMs
>>>> ( ssvm and cpvm ) and virtual router are created fine in the first host,
>>>> and I've succeded to launch VMs in the first host.
>>>>
>>>> But when I tried to launch VMs in the second Host, the VMs always
>>> remains
>>>> unreachable .
>>>> I've checked the traffic connection and I found that the VMs in the
>>> Second
>>>> host can't reach the Virtual router in the First Host , because the VM
>>> and
>>>> Virtual Router are in seperate Link local network. So that why the VMs
>>> in
>>>> the Second Host will not get an IP address after a DHCP discover from
>>> the
>>>> Virtual Router.
>>>>
>>>> My question is : is it a normal behavior that within a cluster , a VM
>>> in a
>>>> Host B can't reach a virtual router in a Host A., did I miss some
>>>> configuration in creating a pool of XenServer ?
>>>>
>>>>
>>>> Best Regards.
>>>>
>>>
>>
>>
> 

Mime
View raw message