#7862 MBS is still firing off F28 builds
Closed: Fixed 4 years ago by mprahl. Opened 4 years ago by ignatenkobrain.

I submitted today another build of ffsend and bat and they were actually built for F28. Since F28 is EOL, it should not do that anymore.

/cc @mprahl @jkaluza


Metadata Update from @smooge:
- Issue assigned to mprahl
- Issue priority set to: Waiting on Assignee (was: Needs Review)
- Issue tagged with: mbs

4 years ago

This is because the F28 platform module is still in the "ready" state.

@mohanboddu do you have a process for retiring modules in the MBS database? If not, I can help you with that.

Edit:
Here is the platform:f28 module:
https://mbs.fedoraproject.org/module-build-service/1/module-builds/1602

@mohanboddu did you see my last comment?

I am expecting that @mohanboddu would see this if this were a releng ticket versus a infrastructure one.

OK turns out mohan doesn't have the ability to do this so I am fixing that. In the meantime

[root@mbs-backend01 ~][PROD]# mbs-manager retire platform:f28
2019-06-17 14:27:13,711 - MainThread - root - INFO - Found 1 module builds:
2019-06-17 14:27:13,711 - MainThread - root - INFO - platform:f28:4:00000000
Retire 1 module builds? [n] y
2019-06-17 14:27:16,778 - MainThread - MBS.models - INFO - <ModuleBuild platform, id=1602, stream=f28, version=4, scratch=None, state 'garbage', batch 0, state_reason 'Module build retired'>, state 5->6
2019-06-17 14:27:16,991 - MainThread - fedmsg.core - DEBUG - Trying to bind to tcp://mbs-backend01.phx2.fedoraproject.org:3000
2019-06-17 14:27:16,992 - MainThread - fedmsg.core - DEBUG - Trying to bind to tcp://mbs-backend01.phx2.fedoraproject.org:3001
2019-06-17 14:27:18,504 - MainThread - root - INFO - Module builds retired.

Needed to run a script from mprahl which got mbooth unstuck

[2019-06-17-12:39] <mprahl> smooge: Yeah, unfortunately, users will be able to submit f28 module builds again. There's nothing to do now. We need to fix the behavior in MBS for a specific use-case before retiring it again. jkaluza has a PR merged that should address most of this, so it shouldn't be too long before we are unblocked on that ticket to retire platform:f28.

Needed to run a script from mprahl which got mbooth unstuck

from module_build_service import conf, db, models

platform = db.session.query(models.ModuleBuild).get(1602)
platform.transition(
    conf, models.BUILD_STATES['ready'], 'Unretired until fm-orchestrator #1243 is resolved')
db.session.add(platform)
db.session.commit()

Just for future reference, to unretire the platform.

Metadata Update from @smooge:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

4 years ago

Metadata Update from @mprahl:
- Issue status updated to: Open (was: Closed)

4 years ago

I've deployed new MBS configuration which should fix this, but we need to add virtual_streams: [fedora] to platform:f28. This will mark the modules built against platform:f28 as usable by other platform:fXY modules.

This can be done by running following script on any MBS prod VM:

from module_build_service import conf, db, models

platform = db.session.query(models.ModuleBuild).get(1602)
platform.update_virtual_streams(db.session, ["fedora"])
db.session.add(platform)
db.session.commit()

We in the end executed following script, because MBS deployed there did not contain the update_virtual_streams method:

from module_build_service import conf, db, models

platform = db.session.query(models.ModuleBuild).get(1602)

vs_obj = db.session.query(models.VirtualStream).filter_by(name="fedora").first()
if vs_obj not in platform.virtual_streams:
    platform.virtual_streams.append(vs_obj)
db.session.add(platform)
db.session.commit()

I've asked @mbooth in https://pagure.io/fm-orchestrator/issue/1243 to check if this issue is fixed now. If he verifies that, we can try retiring platform:f28 once more.

Looks like this is now working? Can we try again? :)

I just retired it by running mbs-manager retire platform:f28 on mbs-backend01.phx2.fedoraproject.org.

Metadata Update from @mprahl:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

4 years ago

Login to comment on this ticket.

Metadata