[Kolla] Few questions about offline deploy & operation.

Eddie Yen missile0407 at gmail.com
Mon May 20 13:35:37 UTC 2019


Hi Mark, glad to see your reply.

1. For python repository, only need are pip and python-docker during
bootstrapping servers. Sounds good.
And I didn't build my own images. I'm using the official binary docker
images.
I'll try to do these on test site.

2. Roger that. Seems little less convenient but still easy to manage by
checking container's status and cluster status for each Openstack services
(like Nova, Neutron, etc.).

3. Okay, I'll try to look inside the container to check about ceph-osd
container.

Anyway, still appreciate your reply. And also apologize my bad English.
Hope this project can be better and more powerful.


Have a nice day,
Eddie.

Mark Goddard <mark at stackhpc.com> 於 2019年5月20日 週一 下午6:32寫道:

> On Fri, 17 May 2019 at 02:46, Eddie Yen <missile0407 at gmail.com> wrote:
>
>> Hi everyone,
>>
>> I'm a newbie of using Kolla-ansible to deploy OpenStack.
>> Everything works well. But I have few questions about Kolla since I was
>> using Fuel as deploy tool before.
>> The questions may silly, but I really want to know since I didn't found
>> much informations to solve my questions..
>>
>>
>>
>> About offline deployment:
>>
>> I already known that we need have OS local repository (CentOS or Ubuntu),
>> docker local package repository, and local docker registry.
>> But one thing I'm not sure that it will use pip to install python
>> packages or not. Because I found it will also install pip into target node
>> during bootstrap-servers.
>> If so, which python package should I prepare for?
>>
>>
> If you are using bootstrap-servers, you will need the pip and docker
> Python packages available for installation.
>
> Are you building your own images? If so you will probably find it easier
> to use binary images than source since these use OS distro packages.
>
>
>> About operation:
>>
>> 1. Is it possible to operate a single service to whole OpenStack nodes?
>> Like using crm to check status about MySQL or RabbitMQ services on all
>> control nodes.
>>
>>
> We don't have a simple way to manage the whole cluster of services from
> one place. You can use the docker CLI, perhaps configured for TCP transport
> if you want remote control, or just via SSH if you don't want to expose it
> to the network.
>
>> 2. How can I maintenance the Ceph OSD if OSD down caused by disk issue?
>> I know how to rebuild OSD by Ceph commands but I'm not sure how to do if
>> ceph-osd running in container.
>>
>>
> I don't tend to use Ceph in kolla, so I can't really answer this one. I
> couldn't see anything about this in the docs.
>
>
>> Many thanks,
>> Eddie.
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20190520/d6d1d8d0/attachment.html>


More information about the openstack-discuss mailing list