You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@cloudstack.apache.org by Ezequiel Mc Govern <ez...@gmail.com> on 2016/11/29 13:40:54 UTC

Change Network Offering from HA to another without HA

In CS 4.8 and in CS 4.9.0.1, changing the NetworkOffering from one with HA to one without HA, the process is performed normally, the redundant VRs are eliminated and a new one is created. Connectivity is maintained and everything works in a normal way. But creating a new VM in the Domain creates a redundant VR, causing connectivity cuts.
Currently we are solving it by performing a manual migration to another Network already configured without HA. Is this by design or can it be considered as a Bug?

Thanks !

In Spanish:

En CS 4.8 y en CS 4.9.0.1 ,al cambiar el NetworkOffering desde uno con HA a uno sin HA, el proceso se efectua normalmente, se eliminan los VR redundantes y se crea uno nuevo. La conectividad se mantiene y funciona todo de forma normal. Pero al crear una nueva VM en el Domain se vuelve a crear un VR redundante provocando cortes en la conectividad.
Actualmente lo estamos solucionando realizando una migración manual a otra Network ya configurado sin HA. ¿ Esto es por diseño o se puede considerar como un Bug ?

Gracias!

Re: Change Network Offering from HA to another without HA

Posted by Dag Sonstebo <Da...@shapeblue.com>.
Nice, glad you got it sorted.

If you have the opportunity it would be very beneficial if you can reproduce this issue in a non-critical network and log a Jira ticket.

Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue

On 30/11/2016, 15:32, "Ezequiel Mc Govern" <ez...@gmail.com> wrote:

    The field “redundant” in table cloud.networks is not modified after the change of the network offering, i modified manualy  the value of "redundant" to false and the problem was resolved :)
    
    
    Thanks!
    
    
    
    > On Nov 29, 2016, at 2:37 PM, Dag Sonstebo <Da...@shapeblue.com> wrote:
    > 
    > Hi Ezequiel,
    > 
    > This sounds like a bug, but I have managed to recreate something similar in the lab. Can you check the following:
    > 
    > First of all find your network in the network table:
    > 
    > SELECT id,name,network_offering_id,state,guest_type,redundant FROM cloud.networks;
    > 
    > The check two things:
    > 1) Check that the network_offering_id is pointed to the correct ID – i.e. the one without HA / redundant VR ( you can check this with SELECT * FROM cloud.network_offerings;).
    > 2) Check that the “redundant” field comes back as zero – “0”
    > 
    > My suspicion is you have the correct network offering in step 1, but you may find the redundant field is set to one  “1” – which would potentially cause the problem you are describing.
    > 
    > I managed to recreate the problem by changing the “redundant” field for an isolated network from 0 to 1 – at which point another VR was created next time I created a VM. I then managed to fix this again by changing the network back to a redundant=0 before restarting the network with cleanup. This deleted both the running VRs and created a single healthy one.
    > 
    > Regards,
    > Dag Sonstebo
    > Cloud Architect
    > ShapeBlue
    > 
    > On 29/11/2016, 13:40, "Ezequiel Mc Govern" <ez...@gmail.com> wrote:
    > 
    >    In CS 4.8 and in CS 4.9.0.1, changing the NetworkOffering from one with HA to one without HA, the process is performed normally, the redundant VRs are eliminated and a new one is created. Connectivity is maintained and everything works in a normal way. But creating a new VM in the Domain creates a redundant VR, causing connectivity cuts.
    >    Currently we are solving it by performing a manual migration to another Network already configured without HA. Is this by design or can it be considered as a Bug?
    > 
    >    Thanks !
    > 
    >    In Spanish:
    > 
    >    En CS 4.8 y en CS 4.9.0.1 ,al cambiar el NetworkOffering desde uno con HA a uno sin HA, el proceso se efectua normalmente, se eliminan los VR redundantes y se crea uno nuevo. La conectividad se mantiene y funciona todo de forma normal. Pero al crear una nueva VM en el Domain se vuelve a crear un VR redundante provocando cortes en la conectividad.
    >    Actualmente lo estamos solucionando realizando una migración manual a otra Network ya configurado sin HA. ¿ Esto es por diseño o se puede considerar como un Bug ?
    > 
    >    Gracias!
    > 
    > 
    > Dag.Sonstebo@shapeblue.com 
    > www.shapeblue.com
    > 53 Chandos Place, Covent Garden, London  WC2N 4HSUK
    > @shapeblue
    > 
    > 
    > 
    
    


Dag.Sonstebo@shapeblue.com 
www.shapeblue.com
53 Chandos Place, Covent Garden, London  WC2N 4HSUK
@shapeblue

Re: Change Network Offering from HA to another without HA

Posted by Ezequiel Mc Govern <ez...@gmail.com>.
The field “redundant” in table cloud.networks is not modified after the change of the network offering, i modified manualy  the value of "redundant" to false and the problem was resolved :)


Thanks!



> On Nov 29, 2016, at 2:37 PM, Dag Sonstebo <Da...@shapeblue.com> wrote:
> 
> Hi Ezequiel,
> 
> This sounds like a bug, but I have managed to recreate something similar in the lab. Can you check the following:
> 
> First of all find your network in the network table:
> 
> SELECT id,name,network_offering_id,state,guest_type,redundant FROM cloud.networks;
> 
> The check two things:
> 1) Check that the network_offering_id is pointed to the correct ID – i.e. the one without HA / redundant VR ( you can check this with SELECT * FROM cloud.network_offerings;).
> 2) Check that the “redundant” field comes back as zero – “0”
> 
> My suspicion is you have the correct network offering in step 1, but you may find the redundant field is set to one  “1” – which would potentially cause the problem you are describing.
> 
> I managed to recreate the problem by changing the “redundant” field for an isolated network from 0 to 1 – at which point another VR was created next time I created a VM. I then managed to fix this again by changing the network back to a redundant=0 before restarting the network with cleanup. This deleted both the running VRs and created a single healthy one.
> 
> Regards,
> Dag Sonstebo
> Cloud Architect
> ShapeBlue
> 
> On 29/11/2016, 13:40, "Ezequiel Mc Govern" <ez...@gmail.com> wrote:
> 
>    In CS 4.8 and in CS 4.9.0.1, changing the NetworkOffering from one with HA to one without HA, the process is performed normally, the redundant VRs are eliminated and a new one is created. Connectivity is maintained and everything works in a normal way. But creating a new VM in the Domain creates a redundant VR, causing connectivity cuts.
>    Currently we are solving it by performing a manual migration to another Network already configured without HA. Is this by design or can it be considered as a Bug?
> 
>    Thanks !
> 
>    In Spanish:
> 
>    En CS 4.8 y en CS 4.9.0.1 ,al cambiar el NetworkOffering desde uno con HA a uno sin HA, el proceso se efectua normalmente, se eliminan los VR redundantes y se crea uno nuevo. La conectividad se mantiene y funciona todo de forma normal. Pero al crear una nueva VM en el Domain se vuelve a crear un VR redundante provocando cortes en la conectividad.
>    Actualmente lo estamos solucionando realizando una migración manual a otra Network ya configurado sin HA. ¿ Esto es por diseño o se puede considerar como un Bug ?
> 
>    Gracias!
> 
> 
> Dag.Sonstebo@shapeblue.com 
> www.shapeblue.com
> 53 Chandos Place, Covent Garden, London  WC2N 4HSUK
> @shapeblue
> 
> 
>

Re: Change Network Offering from HA to another without HA

Posted by Dag Sonstebo <Da...@shapeblue.com>.
Hi Ezequiel,

This sounds like a bug, but I have managed to recreate something similar in the lab. Can you check the following:

First of all find your network in the network table:

SELECT id,name,network_offering_id,state,guest_type,redundant FROM cloud.networks;

The check two things:
1) Check that the network_offering_id is pointed to the correct ID – i.e. the one without HA / redundant VR ( you can check this with SELECT * FROM cloud.network_offerings;).
2) Check that the “redundant” field comes back as zero – “0”

My suspicion is you have the correct network offering in step 1, but you may find the redundant field is set to one  “1” – which would potentially cause the problem you are describing.

I managed to recreate the problem by changing the “redundant” field for an isolated network from 0 to 1 – at which point another VR was created next time I created a VM. I then managed to fix this again by changing the network back to a redundant=0 before restarting the network with cleanup. This deleted both the running VRs and created a single healthy one.

Regards,
Dag Sonstebo
Cloud Architect
ShapeBlue

On 29/11/2016, 13:40, "Ezequiel Mc Govern" <ez...@gmail.com> wrote:

    In CS 4.8 and in CS 4.9.0.1, changing the NetworkOffering from one with HA to one without HA, the process is performed normally, the redundant VRs are eliminated and a new one is created. Connectivity is maintained and everything works in a normal way. But creating a new VM in the Domain creates a redundant VR, causing connectivity cuts.
    Currently we are solving it by performing a manual migration to another Network already configured without HA. Is this by design or can it be considered as a Bug?
    
    Thanks !
    
    In Spanish:
    
    En CS 4.8 y en CS 4.9.0.1 ,al cambiar el NetworkOffering desde uno con HA a uno sin HA, el proceso se efectua normalmente, se eliminan los VR redundantes y se crea uno nuevo. La conectividad se mantiene y funciona todo de forma normal. Pero al crear una nueva VM en el Domain se vuelve a crear un VR redundante provocando cortes en la conectividad.
    Actualmente lo estamos solucionando realizando una migración manual a otra Network ya configurado sin HA. ¿ Esto es por diseño o se puede considerar como un Bug ?
    
    Gracias!


Dag.Sonstebo@shapeblue.com 
www.shapeblue.com
53 Chandos Place, Covent Garden, London  WC2N 4HSUK
@shapeblue