[tripleo][compute][overcloud]Add additional compute nodes on existing overcloud
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* *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.* 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.
On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen@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.
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@redhat.com> wrote:
On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen@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.
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)... Pls advise...Thank you. On Thu, Feb 18, 2021 at 12:14 AM dangerzone ar <dangerzonen@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@redhat.com> wrote:
On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen@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.
Hi, On Thu, Feb 18, 2021 at 4:27 AM dangerzone ar <dangerzonen@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... [2] https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features...
Pls advise...Thank you.
On Thu, Feb 18, 2021 at 12:14 AM dangerzone ar <dangerzonen@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@redhat.com> wrote:
On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen@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.
Hi Martin Noted and thanks. On Thu, Feb 18, 2021 at 3:52 PM Martin Schuppert <mschuppert@redhat.com> wrote:
Hi,
On Thu, Feb 18, 2021 at 4:27 AM dangerzone ar <dangerzonen@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... [2] https://docs.openstack.org/project-deploy-guide/tripleo-docs/latest/features...
Pls advise...Thank you.
On Thu, Feb 18, 2021 at 12:14 AM dangerzone ar <dangerzonen@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@redhat.com> wrote:
On Wed, Feb 17, 2021 at 12:06 PM dangerzone ar <dangerzonen@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.
participants (2)
-
dangerzone ar
-
Martin Schuppert