Fixes #17292 - take otp for freeipa before save @host object in db #5542

pronix · 2018-05-09T17:43:35Z

https://bugzilla.redhat.com/show_bug.cgi?id=1392620

For render user data with otp for freeipa registrytion, host object should have otp before created job for render userdata.

theforeman-bot · 2018-05-09T17:43:36Z

Do not merge! This patch has not been tested yet.

Can an existing organization member please verify this patch?

theforeman-bot · 2018-05-09T17:43:36Z

Do not merge! This patch has not been tested yet.

Can an existing organization member please verify this patch?

theforeman-bot · 2018-05-09T17:43:36Z

Do not merge! This patch has not been tested yet.

Can an existing organization member please verify this patch?

theforeman-bot · 2018-05-09T17:43:44Z

Issues: #17292

theforeman-bot · 2018-05-09T17:43:45Z

There were the following issues with the commit message:

length of the first commit message line for 5896e41 exceeds 65 characters

If you don't have a ticket number, please create an issue in Redmine.

More guidelines are available in Coding Standards or on the Foreman wiki.

This message was auto-generated by Foreman's prprocessor

mmoll · 2018-05-09T18:38:39Z

ok to test

timogoebel

@pronix: Thanks. I left some comments. Do you mind adding a test for this as well? Thanks.

timogoebel · 2018-05-11T09:24:04Z

app/models/concerns/orchestration/realm.rb

@@ -27,6 +27,7 @@ def del_realm

  # Adds the host to the realm, and sets otp if we get one back
  def set_realm(options = {})
+    return self.otp if self.otp.present? && !options[:update]


Why is this needed? Just as a gard, because the realm would be set twice otherwise (once from handle_realm_without_save and once from queue_realm_create)?

it is also big difference when you have record in ipa and trying to update it, and when you trying to recreate token for existed record.

timogoebel · 2018-05-11T09:24:07Z

app/models/concerns/orchestration/compute.rb

@@ -43,6 +43,7 @@ def queue_compute
  end

  def queue_compute_create
+    self.handle_realm_without_save


Isn't it enough to change the priority in queue_realm_create so that this runs earlier than the user data template rendering?

you suggest set priority for queue_realm_create less then 1 ? because setUserData has priority 1

Well, I'd set the priority for the realm update to something lower then setUserdata but try to avoid negative numbers.

so i should shift all tasks priority to +1 and insert on first place queue_realm_create.
correct ?

I think it should be enough to just increase the compute relates tasks, but in general: Yes.

ok. thanks. today will update.

theforeman-bot · 2018-05-24T15:38:32Z

There were the following issues with the commit message:

9e5236a must be in the format fixes #redmine_number - brief description

If you don't have a ticket number, please create an issue in Redmine.

More guidelines are available in Coding Standards or on the Foreman wiki.

This message was auto-generated by Foreman's prprocessor

timogoebel

Looks good to me, but I haven't tested this.

timogoebel · 2018-05-28T06:58:53Z

[test foreman]
[test katello]

pronix · 2018-05-29T10:32:16Z

@timogoebel looks like jenkins error. can you please retrigger it.

mmoll · 2018-05-29T10:43:51Z

test failure unrelated 💚

timogoebel · 2018-05-29T11:25:47Z

Code looks good to me, could somebody please jump in and help with testing this?

eugen-natucci · 2018-05-29T12:15:07Z

@timogoebel help with testing meaning writing tests? I want to participate.

timogoebel · 2018-05-29T12:16:21Z

@jewgeni-prokhorenko: More like taking this patch and verifying it does not break provisioning and fixes the issue by manually testing this.

pronix · 2018-06-14T12:44:44Z

can i help with tests ?

timogoebel · 2018-06-14T12:45:48Z

can i help with tests ?

Sure!

alejojo · 2018-08-08T14:22:36Z

Hi @timogoebel and @pronix,
I just tested with foreman 1.19.0-rc1 katello 3.8 and it fails with the following message:

Render user data template for donna-rosch.minet.net task failed with the following error: undefined method `handle_ca' for # Did you mean? handle_realm

pronix · 2018-08-08T15:01:18Z

hello
i not checked yet with 1.19
during weekend i will recheck it

alejojo · 2018-08-09T08:15:15Z

Hi @pronix,
I've removed the line 99:

 self.handle_ca

in compute.rb and it's working fine :D

mmoll · 2018-09-07T17:45:27Z

@pronix What's the status here? Could you please rebase onto current develop?

pronix · 2018-09-07T17:51:36Z

hello @mmoll
it is in my plan during vacation after 2 weeks :)

schlitzered · 2018-11-22T16:21:26Z

i just tried the patch with my foreman 1.20 installation and it worked, would be great if this could be merged to foreman 1.20.1

mmoll · 2018-11-22T16:40:13Z

a rebase and added tests are still needed.

theforeman-bot · 2018-11-24T14:07:28Z

There were the following issues with the commit message:

e4e69e1 must be in the format fixes #redmine_number - brief description

If you don't have a ticket number, please create an issue in Redmine.

More guidelines are available in Coding Standards or on the Foreman wiki.

This message was auto-generated by Foreman's prprocessor

tinsjourney · 2019-03-06T17:38:38Z

any update on this PR ?
tried the fix on 1.18.0.39 and it worked

timogoebel · 2019-03-07T09:28:15Z

@theforeman/core: Could somebody review this? I would do it, but I don't have a testing environment for realm things.

lzap · 2019-03-07T10:46:10Z

Sorry it took so long. Now, I don't have realm setup as well, but I don't think it's super important to test if the patch fixes the realm issue. What's more important IMHO is test for regressions in compute resource, as you can see this patch changes priority numbers of these and this can have huge impact. I can do that.

Here is an idea - how about to clear priorities for once but forever? The problem is that we have them spread over the codebase and most of them are 1, 2, 3 numbers which makes everything pretty much "random". If you grep queue.create | grep priority you get most of these.

I suggest to create a new file app/services/orchestration/priority.rb with an explicit list of priority numbers as constants:

module Priority
  PRIORITY_CORE_DHCP_CREATE = 10
  PRIORITY_CORE_DHCP_DELETE = 9
  # etc for all orchestration "queue.create" calls

  # Katello
  PRIORITY_KATELLO_CV_CREATE = 30

This could be mixed in everywhere needed and then we would benefit from a single place with all priorities. We could do a breaking change and setup all orchestration order for once and forever so at least in core we have it in expected order and with some reasonable numbers like 100, 200, 300, 310, 320 so plugins could take advantage of that. I would file such a patch if folks are interested - @timogoebel and @tbrisker

Edit: I think it's worth to actually allow plugin authors to commit changes to the file so even plugins cant take advantage of seeing everything in one place. This is not tight coupling, this is just a list of numbers on a single place.

lzap · 2019-03-07T13:08:05Z

A simple example with libvirt VM works just fine:

[app|D|dc6|fe5] Scheduling new DHCP reservations for leigh-canupp.home.lan
[app|D|dc6|fe5] Enqueued task 'dhcp_create_192.168.99.194' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Current organization set to MyOrg
[app|D|dc6|fe5] Current location set to MyLoc
[app|D|dc6|fe5] Current location set to none
[app|D|dc6|fe5] Current organization set to none
[app|D|dc6|fe5] Enqueued task 'Deploy TFTP PXELinux config for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Deploy TFTP PXEGrub2 config for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Deploy TFTP PXEGrub config for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Deploy TFTP iPXE config for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Fetch TFTP boot files for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Set up compute instance leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Enqueued task 'Query instance details for leigh-canupp.home.lan' to 'Host::Managed Main' queue
[app|D|dc6|fe5] Observed create hook on leigh-canupp.home.lan
[app|D|dc6|fe5] Observed postcreate hook on leigh-canupp.home.lan
[app|D|dc6|fe5] Queuing 1 hooks for Host::Managed#postcreate

I am giving this a go, we are unlikely able to test all the plugins.

lzap

Thanks.

lzap · 2019-03-07T13:32:11Z

https://community.theforeman.org/t/idea-one-place-for-orchestration-priorities/13168

theforeman-bot added Waiting on contributor Needs testing Not yet reviewed labels May 9, 2018

pronix force-pushed the user_data_with_ipa_otp branch from 5896e41 to c51dda2 Compare May 9, 2018 17:45

theforeman-bot added Needs testing and removed Waiting on contributor labels May 9, 2018

timogoebel requested changes May 11, 2018

View reviewed changes

theforeman-bot added Waiting on contributor and removed Not yet reviewed labels May 11, 2018

theforeman-bot added Needs re-review Needs testing and removed Waiting on contributor labels May 24, 2018

theforeman-bot added the Waiting on contributor label May 24, 2018

pronix force-pushed the user_data_with_ipa_otp branch from 9e5236a to 54f827e Compare May 24, 2018 15:42

theforeman-bot added Needs re-review Needs testing and removed Waiting on contributor labels May 24, 2018

timogoebel previously approved these changes May 28, 2018

View reviewed changes

theforeman-bot removed the Needs re-review label May 28, 2018

pronix dismissed timogoebel’s stale review via e4e69e1 November 24, 2018 14:07

theforeman-bot added the Waiting on contributor label Nov 24, 2018

Fixes #17292 - take otp before save host

78bb667

pronix force-pushed the user_data_with_ipa_otp branch from e4e69e1 to 78bb667 Compare November 24, 2018 14:13

theforeman-bot added Needs re-review and removed Waiting on contributor labels Nov 24, 2018

lzap approved these changes Mar 7, 2019

View reviewed changes

theforeman-bot removed the Needs re-review label Mar 7, 2019

lzap merged commit b48d799 into theforeman:develop Mar 7, 2019

Fixes #17292 - take otp for freeipa before save @host object in db #5542

Fixes #17292 - take otp for freeipa before save @host object in db #5542

Conversation

pronix commented May 9, 2018

theforeman-bot commented May 9, 2018

theforeman-bot commented May 9, 2018

theforeman-bot commented May 9, 2018

theforeman-bot commented May 9, 2018

theforeman-bot commented May 9, 2018

mmoll commented May 9, 2018

timogoebel left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

theforeman-bot commented May 24, 2018

timogoebel left a comment

Choose a reason for hiding this comment

timogoebel commented May 28, 2018

pronix commented May 29, 2018

mmoll commented May 29, 2018

timogoebel commented May 29, 2018

eugen-natucci commented May 29, 2018

timogoebel commented May 29, 2018

pronix commented Jun 14, 2018

timogoebel commented Jun 14, 2018

alejojo commented Aug 8, 2018

pronix commented Aug 8, 2018

alejojo commented Aug 9, 2018

mmoll commented Sep 7, 2018

pronix commented Sep 7, 2018

schlitzered commented Nov 22, 2018

mmoll commented Nov 22, 2018

theforeman-bot commented Nov 24, 2018

tinsjourney commented Mar 6, 2019 • edited Loading

timogoebel commented Mar 7, 2019

lzap commented Mar 7, 2019 • edited Loading

lzap commented Mar 7, 2019

lzap left a comment

Choose a reason for hiding this comment

lzap commented Mar 7, 2019

tinsjourney commented Mar 6, 2019 •

edited

Loading

lzap commented Mar 7, 2019 •

edited

Loading