[tripleo][overcloud] setup 2 bonds for overcloud deployment tripleo
Hi.. I'm looking help to deploy overcloud (openstack train) as per initial req below:- a) 3 controllers b) 5 computes c) 2 bonds (bond0 - all vlan except for tenant , bond1 - for tenant vlan only) *No external storage and dpdk computes for now as trying to deploy these 3 controllers and 5 computes first. I have gone through openstack web, blogs and some other resources and failed to deploy with success. One of my problems is to identify the right templates to use for my setup and config for 2 bonds above. I have struggled to deploy overcloud for a month without success...I really appreciate your help and support to guide me. Please your help and support is much appreciated. I request help from everybody. Thank you
On Mon, Oct 19, 2020 at 8:14 AM dangerzone ar <dangerzonen@gmail.com> wrote:
Hi.. I'm looking help to deploy overcloud (openstack train) as per initial req below:- a) 3 controllers b) 5 computes c) 2 bonds (bond0 - all vlan except for tenant , bond1 - for tenant vlan only) *No external storage and dpdk computes for now as trying to deploy these 3 controllers and 5 computes first.
I have gone through openstack web, blogs and some other resources and failed to deploy with success. One of my problems is to identify the right templates to use for my setup and config for 2 bonds above. I have struggled to deploy overcloud for a month without success...I really appreciate your help and support to guide me. Please your help and support is much appreciated. I request help from everybody. Thank you
Hello, The network templates in tripleo-heat-templates are always meant as an example which you will need to customize for your specific deployment requirements. That being said, there are a number of examples there which you should be able to adapt to your needs. For 2 bonds with vlans I believe this [1] is the config you will need - nics 2/3 go into one bond then nics 4/5 into another. Since you're looking at train one thing I can think of is to ensure you have the docker-ha [3] environment file included (deployment was by default non-ha [2] before ussuri). Hope it helps you a bit :). Hopefully others can also chime in with more specific info about the network config (I haven't touched those for a very long time) but I think you'll have to give more details about the full deployment command you're currently using (i.e. what are the -e environment_files you're now passing) and the problems you have, regards, marios [1] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... [2] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... [3] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
Hi Marios and team, Thank you for pointing me out. Need further advice. Below is my deployment command :- *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 have copy 2-linux-bonds-vlans to nic-configs folder below:- *cp ~/generated-openstack-tripleo-heat-templates/network/config/2-linux-bonds-vlans/* ~/templates/nic-configs* In the nic-configs folder got 3 files i.README ii.controller.yaml iii.compute-dpdk.yaml My initial deployment only involves controllers and computes (no dpdk), should I use 'compute-dpdk.yaml' file to configure my NIC setup, is it the same template for non-dpdk computes? My actual network, vlan and nic setup:- (1)bond0 (NIC1+NIC2) -Internal API -vlan10 -External -vlan60 -Provisioning - vlan24 *provisoning is configure on the same bond0 of NIC1 and NIC2 (2)bond1 (NIC5+NIC6) -Tenant -vlan20 compute-dpdk.yaml:- The template have nic1 as single nic, while nic2+nic3 on another bond and nic4+nic5 as another bond. Which line should I set as bond0 and bond1? Appreciate to confirm, I would be thankful if someone can point which line I should edit bond0 and bond1 as per setup above. I attached file compute-dpdk.yaml and controller.yaml. Thank you for your help, may god bless you all. On Mon, Oct 19, 2020 at 3:08 PM Marios Andreou <marios@redhat.com> wrote:
On Mon, Oct 19, 2020 at 8:14 AM dangerzone ar <dangerzonen@gmail.com> wrote:
Hi.. I'm looking help to deploy overcloud (openstack train) as per initial req below:- a) 3 controllers b) 5 computes c) 2 bonds (bond0 - all vlan except for tenant , bond1 - for tenant vlan only) *No external storage and dpdk computes for now as trying to deploy these 3 controllers and 5 computes first.
I have gone through openstack web, blogs and some other resources and failed to deploy with success. One of my problems is to identify the right templates to use for my setup and config for 2 bonds above. I have struggled to deploy overcloud for a month without success...I really appreciate your help and support to guide me. Please your help and support is much appreciated. I request help from everybody. Thank you
Hello,
The network templates in tripleo-heat-templates are always meant as an example which you will need to customize for your specific deployment requirements.
That being said, there are a number of examples there which you should be able to adapt to your needs. For 2 bonds with vlans I believe this [1] is the config you will need - nics 2/3 go into one bond then nics 4/5 into another.
Since you're looking at train one thing I can think of is to ensure you have the docker-ha [3] environment file included (deployment was by default non-ha [2] before ussuri).
Hope it helps you a bit :). Hopefully others can also chime in with more specific info about the network config (I haven't touched those for a very long time) but I think you'll have to give more details about the full deployment command you're currently using (i.e. what are the -e environment_files you're now passing) and the problems you have,
regards, marios
[1] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... [2] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... [3] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
Since you don't have a dedicated provisioning port, you want the control plane address directly on the bond rather than on a separate NIC. I modified the templates in 2-linux-bonds-vlans to use the first 2 NICs for the bond with the control plane IP and VLANs, and the other bond for Tenant VLAN (tunnel traffic) and a bridge on the second bond for tenant provider networks. The only changes necessary were moving the IP/subnet/route information from the dedicated NIC to the first bond, and then renumbering the NICs in the bonds to 1/2 and 3/4 instead of 2/3 and 4/5. Note that in order for this example to work, you would need to use LACP bonds with the switch configured with LACP fallback so that the first NIC becomes a regular host port for DHCP and PXE boot prior to the host OS being installed. Once the host OS is installed and the bond is configured, the LACP bond will be negotiated and the provisioning link will become a slave in the first bond. You may want to set the switch to LACP passive so that it waits for the host to bring up the bond. I have attached the modified NIC config templates. -Dan On 10/19/20 5:54 AM, dangerzone ar wrote:
Hi Marios and team, Thank you for pointing me out. Need further advice.
Below is my deployment command :- /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 have copy 2-linux-bonds-vlans to nic-configs folder below:- /cp ~/generated-openstack-tripleo-heat-templates/network/config/2-linux-bonds-vlans/* ~/templates/nic-configs/ / / In the nic-configs folder got 3 files i.README ii.controller.yaml iii.compute-dpdk.yaml
My initial deployment only involves controllers and computes (no dpdk), should I use 'compute-dpdk.yaml' file to configure my NIC setup, is it the same template for non-dpdk computes? My actual network, vlan and nic setup:- (1)bond0 (NIC1+NIC2) -Internal API -vlan10 -External -vlan60 -Provisioning - vlan24 *provisoning is configure on the same bond0 of NIC1 and NIC2
(2)bond1 (NIC5+NIC6) -Tenant -vlan20
compute-dpdk.yaml:- The template have nic1 as single nic, while nic2+nic3 on another bond and nic4+nic5 as another bond. Which line should I set as bond0 and bond1? Appreciate to confirm, I would be thankful if someone can point which line I should edit bond0 and bond1 as per setup above. I attached file compute-dpdk.yaml and controller.yaml.
Thank you for your help, may god bless you all.
On Mon, Oct 19, 2020 at 3:08 PM Marios Andreou <marios@redhat.com <mailto:marios@redhat.com>> wrote:
On Mon, Oct 19, 2020 at 8:14 AM dangerzone ar <dangerzonen@gmail.com <mailto:dangerzonen@gmail.com>> wrote:
Hi.. I'm looking help to deploy overcloud (openstack train) as per initial req below:- a) 3 controllers b) 5 computes c) 2 bonds (bond0 - all vlan except for tenant , bond1 - for tenant vlan only) *No external storage and dpdk computes for now as trying to deploy these 3 controllers and 5 computes first.
I have gone through openstack web, blogs and some other resources and failed to deploy with success. One of my problems is to identify the right templates to use for my setup and config for 2 bonds above. I have struggled to deploy overcloud for a month without success...I really appreciate your help and support to guide me. Please your help and support is much appreciated. I request help from everybody. Thank you
Hello,
The network templates in tripleo-heat-templates are always meant as an example which you will need to customize for your specific deployment requirements.
That being said, there are a number of examples there which you should be able to adapt to your needs. For 2 bonds with vlans I believe this [1] is the config you will need - nics 2/3 go into one bond then nics 4/5 into another.
Since you're looking at train one thing I can think of is to ensure you have the docker-ha [3] environment file included (deployment was by default non-ha [2] before ussuri).
Hope it helps you a bit :). Hopefully others can also chime in with more specific info about the network config (I haven't touched those for a very long time) but I think you'll have to give more details about the full deployment command you're currently using (i.e. what are the -e environment_files you're now passing) and the problems you have,
regards, marios
[1] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... <https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5f0ac09712c98fe8560d752544f80/network/config/2-linux-bonds-vlans/role.role.j2.yaml#L169-L188> [2] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... <https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5f0ac09712c98fe8560d752544f80/overcloud-resource-registry-puppet.j2.yaml#L172> [3] https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5... <https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5f0ac09712c98fe8560d752544f80/environments/docker-ha.yaml>
-- Dan Sneddon | Senior Principal Software Engineer dsneddon@redhat.com | redhat.com/cloud dsneddon:irc | @dxs:twitter
Hi Dan..noted and thank you. I will proceed wth the deployment. Thank you so much for your great support. Thanks to all too.. Have a nice day... On Tue, 20 Oct 2020 23:24 Dan Sneddon, <dsneddon@redhat.com> wrote:
Since you don't have a dedicated provisioning port, you want the control plane address directly on the bond rather than on a separate NIC. I modified the templates in 2-linux-bonds-vlans to use the first 2 NICs for the bond with the control plane IP and VLANs, and the other bond for Tenant VLAN (tunnel traffic) and a bridge on the second bond for tenant provider networks.
The only changes necessary were moving the IP/subnet/route information from the dedicated NIC to the first bond, and then renumbering the NICs in the bonds to 1/2 and 3/4 instead of 2/3 and 4/5.
Note that in order for this example to work, you would need to use LACP bonds with the switch configured with LACP fallback so that the first NIC becomes a regular host port for DHCP and PXE boot prior to the host OS being installed. Once the host OS is installed and the bond is configured, the LACP bond will be negotiated and the provisioning link will become a slave in the first bond. You may want to set the switch to LACP passive so that it waits for the host to bring up the bond.
I have attached the modified NIC config templates.
-Dan
On 10/19/20 5:54 AM, dangerzone ar wrote:
Hi Marios and team, Thank you for pointing me out. Need further advice.
Below is my deployment command :- /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 have copy 2-linux-bonds-vlans to nic-configs folder below:- /cp
~/generated-openstack-tripleo-heat-templates/network/config/2-linux-bonds-vlans/*
~/templates/nic-configs/ / / In the nic-configs folder got 3 files i.README ii.controller.yaml iii.compute-dpdk.yaml
My initial deployment only involves controllers and computes (no dpdk), should I use 'compute-dpdk.yaml' file to configure my NIC setup, is it the same template for non-dpdk computes? My actual network, vlan and nic setup:- (1)bond0 (NIC1+NIC2) -Internal API -vlan10 -External -vlan60 -Provisioning - vlan24 *provisoning is configure on the same bond0 of NIC1 and NIC2
(2)bond1 (NIC5+NIC6) -Tenant -vlan20
compute-dpdk.yaml:- The template have nic1 as single nic, while nic2+nic3 on another bond and nic4+nic5 as another bond. Which line should I set as bond0 and bond1? Appreciate to confirm, I would be thankful if someone can point which line I should edit bond0 and bond1 as per setup above. I attached file compute-dpdk.yaml and controller.yaml.
Thank you for your help, may god bless you all.
On Mon, Oct 19, 2020 at 3:08 PM Marios Andreou <marios@redhat.com <mailto:marios@redhat.com>> wrote:
On Mon, Oct 19, 2020 at 8:14 AM dangerzone ar <dangerzonen@gmail.com <mailto:dangerzonen@gmail.com>> wrote:
Hi.. I'm looking help to deploy overcloud (openstack train) as per initial req below:- a) 3 controllers b) 5 computes c) 2 bonds (bond0 - all vlan except for tenant , bond1 - for tenant vlan only) *No external storage and dpdk computes for now as trying to deploy these 3 controllers and 5 computes first.
I have gone through openstack web, blogs and some other resources and failed to deploy with success. One of my problems is to identify the right templates to use for my setup and config for 2 bonds above. I have struggled to deploy overcloud for a month without success...I really appreciate your help and support to guide me. Please your help and support is much appreciated. I request help from everybody. Thank you
Hello,
The network templates in tripleo-heat-templates are always meant as an example which you will need to customize for your specific deployment requirements.
That being said, there are a number of examples there which you should be able to adapt to your needs. For 2 bonds with vlans I believe this [1] is the config you will need - nics 2/3 go into one bond then nics 4/5 into another.
Since you're looking at train one thing I can think of is to ensure you have the docker-ha [3] environment file included (deployment was by default non-ha [2] before ussuri).
Hope it helps you a bit :). Hopefully others can also chime in with more specific info about the network config (I haven't touched those for a very long time) but I think you'll have to give more details about the full deployment command you're currently using (i.e. what are the -e environment_files you're now passing) and the problems you have,
regards, marios
[1]
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
<
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
[2]
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
<
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
[3]
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
<
https://opendev.org/openstack/tripleo-heat-templates/src/commit/ed46b26527b5...
-- Dan Sneddon | Senior Principal Software Engineer dsneddon@redhat.com | redhat.com/cloud dsneddon:irc | @dxs:twitter
participants (3)
-
Dan Sneddon
-
dangerzone ar
-
Marios Andreou