Bug 577796 - RFE: Support installing task dependencies only after kickstart is already complete
Summary: RFE: Support installing task dependencies only after kickstart is already com...
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Beaker
Classification: Retired
Component: beah
Version: 0.5
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: beaker-dev-list
QA Contact:
URL:
Whiteboard: SimpleHarness
Depends On:
Blocks: 472740 496584 651828
TreeView+ depends on / blocked
 
Reported: 2010-03-29 09:23 UTC by Alexander Todorov
Modified: 2019-01-21 17:42 UTC (History)
8 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2019-01-21 17:42:32 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 554844 0 low CLOSED Document how to bypass the task library 2021-02-22 00:41:40 UTC

Description Alexander Todorov 2010-03-29 09:23:01 UTC
Description of problem:
Currently there are some additional packages/package groups which are added into the %packages section and are installed from the installer. See bug #192984 which requests a mechanism to do so. 

This RFE requests that those additional packages are not included in the %packages section but installed via %post or after the system reboots (preferred) prior to start of test cases using yum. 

This will give snake-workflow 100% control of what packages are installed by Anaconda (listed in install.log) and will make it possible for RTT to execute the default-package-set test case. Additional packages installed after system reboot is preferred because it will not interfere with RTT's test cases.

Comment 1 Bill Peck 2010-04-07 12:35:59 UTC
Notice:

Legacy RHTS is soon to be retired and replaced by Beaker. As part of
this migration process all RHTS bugs need to be re-verified against a
Beaker instance by the cut-off date

   5pm Monday April 12th UTC-4. 

Please confirm this bug is still relevant to Beaker by re-verifying it
against the stage deployment of Beaker https://beaker-stage.app.eng.bos.redhat.com.

To keep this bug open please comment on it

If it has not received a comment by that date the bug will be closed/wontfix.

After the cutoff date all commented bugs will be moved to the Beaker
product.


thank you

Comment 2 Alexander Todorov 2010-04-08 12:47:21 UTC
Hi,
this is still an issue:
https://beaker-stage.app.eng.bos.redhat.com/recipes/436#task1843

I propose to merge with bug #472740 and implement the fix for both of them.

Comment 3 Marian Csontos 2010-08-06 09:41:39 UTC
We need to re-implement dependencies on harness side, to handle e.g. conflicting dependencies for tasks in single recipe...

This should be part of larger task packaging rework - see Bug 554844.

Comment 4 David Kutálek 2010-11-10 13:10:47 UTC
(In reply to comment #3)
> We need to re-implement dependencies on harness side, to handle e.g.
> conflicting dependencies for tasks in single recipe...
> 
> This should be part of larger task packaging rework - see Bug 554844.

So you propose harness should be able to uninstall conflicting packages when needed?

Comment 5 Marian Csontos 2010-11-11 07:34:10 UTC
(In reply to comment #4)
Yes, it has to.

Comment 6 Nick Coghlan 2012-10-17 04:35:31 UTC
Bulk reassignment of issues as Bill has moved to another team.

Comment 7 Nick Coghlan 2012-11-07 05:58:47 UTC
Reverting to NEW to more accurately reflect current status.

Comment 9 Nick Coghlan 2014-08-06 08:03:11 UTC
Alexander, is this still needed?

workflow-snake is going away, and "remote_post" and "ks_appends" allow arbitrary code execution in %post

Setting the "packages" ks_meta variable also completely overrides the default templates.

Comment 10 Alexander Todorov 2014-08-06 09:24:55 UTC
(In reply to Nick Coghlan from comment #9)
> Alexander, is this still needed?
> 
> workflow-snake is going away, and "remote_post" and "ks_appends" allow
> arbitrary code execution in %post
> 
> Setting the "packages" ks_meta variable also completely overrides the
> default templates.

Hi Nick,
does Beaker override the package selection specified in the <kickstart> tag ? This is the primary way QE specifies system and package configuration. If not changed then I think this request may be closed.

Comment 11 Nick Coghlan 2014-08-12 06:58:47 UTC
Dan, could you take a look at Alexander's Q in comment 10? That will determine what, if anything, still needs to be done for this RFE.

Comment 12 Dan Callaghan 2014-08-12 07:06:31 UTC
(In reply to Alexander Todorov from comment #10)
> does Beaker override the package selection specified in the <kickstart> tag
> ? This is the primary way QE specifies system and package configuration. If
> not changed then I think this request may be closed.

If you have a recipe with

<kickstart>
...
%packages
some-package
...
</kickstart>

then Beaker uses your %packages section as given, but also appends any other packages required by the tasks in the recipe.

As noted by Nick you can pass packages= kickstart metadata, but then you are on your own in terms of satisfying your tasks' requirements.

I guess the original point of this RFE was to make Beaker stop using %packages to install task requirements, and to make the harness do the installation instead. That request remains unfulfilled. We are halfway there with restraint, which can install task requirements, but Beaker still always munges %packages to insert the task requirements there too (unless all the tasks in your recipe are "external" tasks).

Comment 13 Alexander Todorov 2014-08-14 07:10:51 UTC
(In reply to Dan Callaghan from comment #12)

> I guess the original point of this RFE was to make Beaker stop using
> %packages to install task requirements, and to make the harness do the
> installation instead. That request remains unfulfilled. We are halfway there
> with restraint, 


I also think this was the original point behind this request. Do you have a tracker bug for this wrt restraint or not? We can use the current one to track progress or I can open another one and close this one if you prefer.

Comment 14 Nick Coghlan 2014-08-18 04:03:16 UTC
Keeping this one open is fine - I just changed the title to more clearly express the intent.

Perhaps a ks_meta flag would suffice?

Comment 15 Dan Callaghan 2017-01-31 16:11:58 UTC
This is implemented in restraint, should it be considered as WONTFIX for beah?


Note You need to log in before you can comment on or make changes to this bug.