Open Stack

Hi Martin
Noted and thanks.




On Thu, Feb 18, 2021 at 3:52 PM Martin Schuppert <mschuppert at redhat.com>
wrote:

>
> Hi,
>
> On Thu, Feb 18, 2021 at 4:27 AM dangerzone ar <dangerzonen at gmail.com>
> wrote:
>
>> Hi Martin and Team,
>> Following to my questions of adding new compute onto existing overcloud,
>> If the new compute server is not the same type (existing running on blade
>> server) where it may be rack mount type, what are the changes I should make
>> to all yaml files related? I foreseen tricky especially the nic layout is
>> different between blade and rack mount type.
>> Pls advise anyone facing the same scenario.... (I just been informed the
>> new compute server is rack mount type)...
>>
>
> If you have different kind of compute nodes you should create a new
> compute role for the new type [1].
> You can start with making a copy of your current compute role section in
> the roles_data.yaml file and
> name it to identify the new ones. Then create a network configuration for
> this role and register it in
> your parameter files, like you have for you current compute role.
>
> With the new role you can also configure different parameters for e.g. not
> depending on the hardware
> and usage [2].
>
> HTH,
> Martin
>
> [1]
> https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features/custom_roles.html
> [2]
> https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features/role_specific_parameters.html
>
>
>
>> Pls advise...Thank you.
>>
>> On Thu, Feb 18, 2021 at 12:14 AM dangerzone ar <dangerzonen at gmail.com>
>> wrote:
>>
>>> Hi Martin,
>>> Noted and thanks for your response and input.
>>> If I can remember, ComputeCount can be defined in a node-info file...
>>> Thus, If I edit ComputeCount from existing 3 to 6 in node-info file...I
>>> can run the same overcloud deploy command as follows
>>> ```
>>>   openstack overcloud deploy --templates \
>>>   -r ~/templates/roles_data.yaml \
>>>   -n ~/templates/network_data.yaml \
>>>   -e
>>> /usr/share/openstack-tripleo-heat-templates/environments/network-environment.yaml
>>> \
>>>   -e
>>> /usr/share/openstack-tripleo-heat-templates/environments/network-isolation.yaml
>>> \
>>>   -e
>>> /usr/share/openstack-tripleo-heat-templates/environments/docker-ha.yaml
>>>   -e ~/templates/network.yaml \
>>>   -e ~/templates/scheduler_hints_env.yaml \
>>>   -e ~/templates/*node-info.yaml*
>>> ```
>>> I hope this is good to go....
>>> This scaling new compute node should not affect existing openstack nodes
>>> and instances running right...as I understand...redeployment overcloud..
>>> just updates and registers new 3 compute nodes onto the overcloud...... I'm
>>> a bit nervous...so I need to plan well and identify any risks...
>>>
>>> Thanks Martin and team...
>>>
>>> On Wed, Feb 17, 2021 at 7:18 PM Martin Schuppert <mschuppert at redhat.com>
>>> wrote:
>>>
>>>>
>>>> On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen at gmail.com> wrote:
>>>>
>>>> Hi all...I'm planning to add another 3 compute servers to existing
>>>>> overcloud (consists of existing 2 compute nodes, 3 controllers and 3
>>>>> storage). I'm running on openstack ussuri using tripleo and centos7.
>>>>>
>>>>> Below is the command and parameters used during initial deployment
>>>>> ```
>>>>> openstack overcloud deploy --templates \
>>>>>   -r ~/templates/roles_data.yaml \
>>>>>   -n ~/templates/network_data.yaml \
>>>>>   -e
>>>>> /usr/share/openstack-tripleo-heat-templates/environments/network-environment.yaml
>>>>> \
>>>>>   -e
>>>>> /usr/share/openstack-tripleo-heat-templates/environments/network-isolation.yaml
>>>>> \
>>>>>   -e
>>>>> /usr/share/openstack-tripleo-heat-templates/environments/docker-ha.yaml
>>>>>   -e ~/templates/network.yaml \
>>>>>   -e ~/templates/scheduler_hints_env.yaml \
>>>>>   -e ~/templates/node-info.yaml
>>>>> ```
>>>>> So now this is my plan to add additional compute nodes
>>>>> *1. Create new json file containing details of new compute nodes i.e.
>>>>> newnodes.json*
>>>>> *openstack overcloud node import newnodes.json*
>>>>> *2. Run introspection*
>>>>> *openstack overcloud node introspect --all-manageable*
>>>>> *3. Tag each new node for its desired role*
>>>>> *openstack baremetal node set --property
>>>>> capabilities='profile:compute,boot_option:local' [NODE UUID]*
>>>>> *4. Run openstack overcloud deploy*
>>>>> *openstack overcloud deploy --compute-scale 3*
>>>>>
>>>>
>>>> The scale parameters, like --compute-scale, are deprecated. You should
>>>> just change the <Role>Count (e.g. ComputeCount)
>>>> to match your new scale. At the moment you have 3, change it to 6.
>>>>
>>>>
>>>>> *Step No 4, should I run overcloud deploy with all the environment
>>>>> template/parameters same as during initial deployment as per example above?*
>>>>> *Really appreciate your advice of these steps especially no 4.*
>>>>>
>>>>
>>>> yes, run the same deploy command with the modified <Role>Count.
>>>>
>>>>
>>>>>
>>>>> Please advise further if my plan and steps above are good to go.
>>>>> Thank you for your help and kind advice. May God bless you.
>>>>>
>>>>> Thank you.
>>>>> dz
>>>>>
>>>>> 1.
>>>>>
>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20210218/caf80d19/attachment-0001.html>