Cannot deploly an instance at a specific host --(no such host found) [nova]

Eugen Block eblock at nde.ag
Fri Nov 13 08:17:21 UTC 2020


The database connection doesn't look right, the compute node also  
needs access to the nova database (localhost is not reachable for  
another compute node) and usually with the nova user, not root, but  
that might be a devstack thing, I can't tell.
The db settings in your nova.conf also seem wrong:

[database]
connection = sqlite:////var/lib/nova/nova.sqlite

This should point to the nova database on the control node matching  
the one from 'nova-manage cell_v2 list_cells' command.

I think there are lots of things not matching a multi-node setup,  
you'll need to fix that. As I said, I'm not familiar with devstack at  
all so I can't really tell exactly what needs to be done.

Regards,
Eugen


Zitat von Pavlos Basaras <pbasaras at gmail.com>:

> Hello,
>
> thanks for your help.
>
> I only added one extra compute node, and i changed the nova.conf on this
> node to match the controller's.
>
> Here are the commands.
>
> $ nova-manage cell_v2 list_cells --verbose
> +-------+--------------------------------------+-----------------------------------------------------------+--------------------------------------------------------------+----------+
> |  Name |                 UUID                 |
> Transport URL                       |                     Database
> Connection                      | Disabled |
> +-------+--------------------------------------+-----------------------------------------------------------+--------------------------------------------------------------+----------+
> | cell0 | 00000000-0000-0000-0000-000000000000 |
>  none:///                         | mysql+pymysql://
> root:linux at 127.0.0.1/nova_cell0?charset=utf8 |  False   |
> | cell1 | 1522c22f-64d4-4882-8ae8-ed0f9407407c | rabbit://
> stackrabbit:linux at 192.168.40.184:5672/nova_cell1 | mysql+pymysql://
> root:linux at 127.0.0.1/nova_cell1?charset=utf8 |  False   |
> +-------+--------------------------------------+-----------------------------------------------------------+--------------------------------------------------------------+----------+
>
> $ sudo  rabbitmqctl list_vhosts
> Listing vhosts
> /
> nova_cell1
>
>
> $ sudo  rabbitmqctl list_permissions -p nova_cell1
> Listing permissions in vhost "nova_cell1"
> stackrabbit .* .* .*
>
>
> best,
> P.
>
>
> On Thu, Nov 12, 2020 at 11:16 AM Eugen Block <eblock at nde.ag> wrote:
>
>> Did you change the transport_url only in the nova.conf of the new
>> compute node? Or did you also change the rabbitmq configuration? The
>> transport_url should match the actual rabbitmq config, of course, and
>> it should be the same on every node.
>> Does that match your cell setup? To compare run:
>>
>> control:~ # nova-manage cell_v2 list_cells --verbose
>>
>> Do you have a vhost configuration? (I'm not familiar with devstack)
>> Here's an example of a "nova" vhost with respective permissions:
>>
>>
>> control:~ # rabbitmqctl list_vhosts
>> Listing vhosts
>> /
>> nova
>>
>> control:~ # rabbitmqctl list_permissions -p nova
>> Listing permissions in vhost "nova"
>> nova    .*      .*      .*
>>
>>
>> If there's a vhost it should be also reflected in the transport_url.
>> Maybe you should share all of the above information here.
>>
>>
>> Zitat von Pavlos Basaras <pbasaras at gmail.com>:
>>
>> > Hello,
>> >
>> > maybe there is stg wrong with the installation.
>> >
>> > Let me clarify a few things.
>> >
>> > I have devstack deployed in a VM (pre-installed: keystone, glance, nova,
>> > placement, cinder, neutron, and horizon.)
>> > I can deploy machines successfully at the devstack controller space
>> > --everything seems to work fine. --> is seems to work fine with
>> opensource
>> > mano as well
>> >
>> > I am trying to add another pc as a compute host to be able to deploy vms
>> at
>> > this new compute host following (
>> >
>> https://docs.openstack.org/nova/queens/install/compute-install-ubuntu.html
>> )
>> >
>> > Also attached the nova.conf file. The only major differences are:
>> >  --the transport url for rabbitmq which i made according to the transport
>> > url of the controller, i.e., instead of
>> rabbit://openstack:linux@controller
>> >  i have   rabbit://stackrabbit:linux@controller
>> > -- i replaced the ports with the service, e.g., instead using the 5000
>> port
>> > --> "http://controller/identity/v3" instead of  "
>> http://controller:5000/v3"
>> >
>> > please excuse (any) newbie questions
>> >
>> > all the best,
>> > Pavlos.
>> >
>> >
>> > On Wed, Nov 11, 2020 at 4:03 PM Eugen Block <eblock at nde.ag> wrote:
>> >
>> >> There might be some mixup during the setup, I'm not sure how the other
>> >> cell would be created. I'd probably delete the cell with UUID
>> >> 1a0fde85-8906-46fb-b721-01a28c978439 and retry the discover_hosts with
>> >> the right cell UUID:
>> >>
>> >> nova-manage cell_v2 delete_cell --cell_uuid
>> >> 1a0fde85-8906-46fb-b721-01a28c978439
>> >> nova-manage cell_v2 discover_hosts --cell_uuid
>> >> 1522c22f-64d4-4882-8ae8-ed0f9407407c
>> >>
>> >> Does that work?
>> >>
>> >>
>> >> Zitat von Pavlos Basaras <pbasaras at gmail.com>:
>> >>
>> >> > Hello,
>> >> >
>> >> > yes i have this discover_hosts_in_cells_interval = 300
>> >> >
>> >> >
>> >> > interestingly when i issued a combination of map_cell_and_hosts and
>> >> > update_cell
>> >> >
>> >> > the output from: nova-manage cell_v2 list_hosts
>> >> > +-----------+--------------------------------------+-------------+
>> >> > | Cell Name |              Cell UUID               |   Hostname  |
>> >> > +-----------+--------------------------------------+-------------+
>> >> > |    None   | 1a0fde85-8906-46fb-b721-01a28c978439 | computenode |
>> >> > |    None   | 1a0fde85-8906-46fb-b721-01a28c978439 |     nrUE    |
>> >> > |   cell1   | 1522c22f-64d4-4882-8ae8-ed0f9407407c |  openstack  |
>> >> > +-----------+--------------------------------------+-------------+
>> >> >
>> >> > when the new compute nodes seem to not have a cell mapped
>> >> >
>> >> > best,
>> >> > P.
>> >> >
>> >> >
>> >> > On Wed, Nov 11, 2020 at 3:26 PM Eugen Block <eblock at nde.ag> wrote:
>> >> >
>> >> >> Hm,
>> >> >>
>> >> >> indeed rather strange to me.
>> >> >>
>> >> >> Do you see anything in the nova-scheduler.log? If you activated the
>> >> >>
>> >> >> discover_hosts_in_cells_interval = 300
>> >> >>
>> >> >> it should query for new hosts every 5 minutes.
>> >> >>
>> >> >>
>> >> >>
>> >> >> Zitat von Pavlos Basaras <pbasaras at gmail.com>:
>> >> >>
>> >> >> > Hello,
>> >> >> >
>> >> >> > thanks very much for your prompt reply.
>> >> >> >
>> >> >> >
>> >> >> > regarding the first command "nova-manage cell_v2 list_hosts" the
>> >> output
>> >> >> is
>> >> >> > the following (openstack is the host of the controller). I dont see
>> >> any
>> >> >> > other node here, even when i execute  the discover_hosts command
>> >> >> >
>> >> >> > +-----------+--------------------------------------+-----------+
>> >> >> > | Cell Name |              Cell UUID               |  Hostname |
>> >> >> > +-----------+--------------------------------------+-----------+
>> >> >> > |   cell1   | 1522c22f-64d4-4882-8ae8-ed0f9407407c | openstack |
>> >> >> > +-----------+--------------------------------------+-----------+
>> >> >> >
>> >> >> >
>> >> >> > Also this is the output from my controller when i use the command:
>> >> sudo
>> >> >> -s
>> >> >> > /bin/sh -c "nova-manage cell_v2 discover_hosts --verbose" nova
>>  (not
>> >> >> sure
>> >> >> > if this helps)
>> >> >> > Found 2 cell mappings.
>> >> >> > Skipping cell0 since it does not contain hosts.
>> >> >> > Getting computes from cell 'cell1':
>> >> 1522c22f-64d4-4882-8ae8-ed0f9407407c
>> >> >> > Found 0 unmapped computes in cell:
>> >> 1522c22f-64d4-4882-8ae8-ed0f9407407c
>> >> >> >
>> >> >> > any thoughts?
>> >> >> >
>> >> >> >
>> >> >> > best,
>> >> >> > Pavlos.
>> >> >>
>> >> >>
>> >> >>
>> >> >>
>> >> >>
>> >>
>> >>
>> >>
>> >>
>>
>>
>>
>>






More information about the openstack-discuss mailing list