[ironic] A new project for useful deploy steps?
Hi all, We have finally implemented in-band deploy steps (w00t!), and people started coming up with ideas. I have two currently: 1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init) I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a collection of potentially useful deploy steps? Or should we nonetheless just put them in IPA with priority=0? Opinions welcome. Dmitry -- Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
Hi Dmitry, Thanks for bringing this up! We discussed this in our weekly ironic meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users. Would be good to hear from others, if they don't think this is a good idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :) --ruby [1] http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron... On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com> wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people started coming up with ideas. I have two currently: 1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a collection of potentially useful deploy steps?
Or should we nonetheless just put them in IPA with priority=0?
Opinions welcome.
Dmitry
-- Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
I concur. The only thing with RFE's and patch approvals is I think we should remember that we want it to be easy. So processes like RFEs may not be helpful to a "oh, this tiny little thing makes a lot of sense" sort of things, since it quickly becomes a situation where you spend more time on the RFE than the patch itself. On Mon, Feb 8, 2021 at 7:43 AM Ruby Loo <opensrloo@gmail.com> wrote:
Hi Dmitry,
Thanks for bringing this up! We discussed this in our weekly ironic meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users.
Would be good to hear from others, if they don't think this is a good idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :)
--ruby
[1] http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron...
On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com> wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people started coming up with ideas. I have two currently: 1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a collection of potentially useful deploy steps?
Or should we nonetheless just put them in IPA with priority=0?
Opinions welcome.
Dmitry
-- Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
So true. Although... On Mon, Feb 8, 2021 at 11:09 AM Julia Kreger <juliaashleykreger@gmail.com> wrote:
I concur. The only thing with RFE's and patch approvals is I think we should remember that we want it to be easy. So processes like RFEs may not be helpful to a "oh, this tiny little thing makes a lot of sense" sort of things, since it quickly becomes a situation where you spend more time on the RFE than the patch itself.
On Mon, Feb 8, 2021 at 7:43 AM Ruby Loo <opensrloo@gmail.com> wrote:
Hi Dmitry,
Thanks for bringing this up! We discussed this in our weekly ironic
meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users.
Would be good to hear from others, if they don't think this is a good
idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :)
--ruby
[1]
http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron...
On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com>
wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people
started coming up with ideas. I have two currently:
1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a
collection of potentially useful deploy steps? >> >> Or should we nonetheless just put them in IPA with priority=0? >> >> Opinions welcome. >> >> Dmitry >> >> -- >> Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, >> Commercial register: Amtsgericht Muenchen, HRB 153243, >> Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
If it is something small, wouldn't it still require a short description in a story. With '[RFE]' or something like that in the title? (Or have I totally forgotten how we do things). --ruby On Fri, Feb 12, 2021 at 6:00 PM Ruby Loo <opensrloo@gmail.com> wrote:
So true. Although...
On Mon, Feb 8, 2021 at 11:09 AM Julia Kreger <juliaashleykreger@gmail.com> wrote:
I concur. The only thing with RFE's and patch approvals is I think we should remember that we want it to be easy. So processes like RFEs may not be helpful to a "oh, this tiny little thing makes a lot of sense" sort of things, since it quickly becomes a situation where you spend more time on the RFE than the patch itself.
On Mon, Feb 8, 2021 at 7:43 AM Ruby Loo <opensrloo@gmail.com> wrote:
Hi Dmitry,
Thanks for bringing this up! We discussed this in our weekly ironic
meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users.
Would be good to hear from others, if they don't think this is a good
idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :)
--ruby
[1]
http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron...
On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com>
wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people
started coming up with ideas. I have two currently:
1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a
collection of potentially useful deploy steps? >> >> Or should we nonetheless just put them in IPA with priority=0? >> >> Opinions welcome. >> >> Dmitry >> >> -- >> Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, >> Commercial register: Amtsgericht Muenchen, HRB 153243, >> Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
I absolutely agree that we should still file RFE stories, at least for tracking of what was done, when and why. On Sat, Feb 13, 2021 at 12:03 AM Ruby Loo <opensrloo@gmail.com> wrote:
If it is something small, wouldn't it still require a short description in a story. With '[RFE]' or something like that in the title? (Or have I totally forgotten how we do things).
--ruby
On Fri, Feb 12, 2021 at 6:00 PM Ruby Loo <opensrloo@gmail.com> wrote:
So true. Although...
On Mon, Feb 8, 2021 at 11:09 AM Julia Kreger <juliaashleykreger@gmail.com> wrote:
I concur. The only thing with RFE's and patch approvals is I think we should remember that we want it to be easy. So processes like RFEs may not be helpful to a "oh, this tiny little thing makes a lot of sense" sort of things, since it quickly becomes a situation where you spend more time on the RFE than the patch itself.
On Mon, Feb 8, 2021 at 7:43 AM Ruby Loo <opensrloo@gmail.com> wrote:
Hi Dmitry,
Thanks for bringing this up! We discussed this in our weekly ironic
meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users.
Would be good to hear from others, if they don't think this is a good
idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :)
--ruby
[1]
http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron...
On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com>
wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people
started coming up with ideas. I have two currently:
1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a
collection of potentially useful deploy steps? >> >> Or should we nonetheless just put them in IPA with priority=0? >> >> Opinions welcome. >> >> Dmitry >> >> -- >> Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, >> Commercial register: Amtsgericht Muenchen, HRB 153243, >> Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
-- Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
On Mon, 8 Feb 2021 at 15:36, Ruby Loo <opensrloo@gmail.com> wrote:
Hi Dmitry,
Thanks for bringing this up! We discussed this in our weekly ironic meeting [1]. The consensus there seems to be to keep the ideas in IPA (with priority=0). The additional code will be 'negligible' in size so ramdisk won't be bloated due to this. Also, it keeps things simple. Having a separate package means more maintenance overhead and confusion for our users.
Would be good to hear from others, if they don't think this is a good idea. Otherwise, I'm looking forward to Dmitry's RFEs on this :)
--ruby
[1] http://eavesdrop.openstack.org/irclogs/%23openstack-ironic/%23openstack-iron...
On Mon, Feb 8, 2021 at 8:02 AM Dmitry Tantsur <dtantsur@redhat.com> wrote:
Hi all,
We have finally implemented in-band deploy steps (w00t!), and people started coming up with ideas. I have two currently: 1) configure arbitrary kernel command line arguments via grub 2) write NetworkManager configuration (for those not using cloud-init)
Seems like a good idea. I'd love to see deploy steps/templates getting more traction.
I'm not sure how I feel about putting these in IPA proper, seems like we may go down a rabbit hole here. But what about a new project (ironic-python-agent-extras?) with a hardware manager providing a collection of potentially useful deploy steps?
Or should we nonetheless just put them in IPA with priority=0?
Seems reasonable to me, if they are generic enough.
Opinions welcome.
Dmitry
-- Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill
participants (4)
-
Dmitry Tantsur
-
Julia Kreger
-
Mark Goddard
-
Ruby Loo