#248 Deploy rest of artifacts bits
Closed: Fixed None Opened 8 years ago by tflink.

I started this to track what all needs to be done to get the recent changes to artifacts deployed. The stuff I know about:
* build and release new libtaskotron
* update dev/stg to current develop branch for depcheck, upgradepath
* deploy new libtaskotron to dev/stg

Assuming all goes well:
* update libtaskotron on prod
* merge develop branch into master for depcheck, upgradepath

If I'm forgetting something, please add it to the description or in a comment


One further task: In ansible, we need to configure taskotron.yaml with the new keys: artifacts_baseurl and resultsdb_frontend

I added #520 to block #399, but it does not prevent dev/stg deployment, it's just something to keep an eye on.

Another thing to check - make sure that we really use develop branch on dev/stg for all our tasks, and master for production. Because I think we've done some specific hacks about this in the past, but I'm not sure it is really consistent for all our tasks. I have created develop branch for rpmlint just now, it hasn't existed before.

Note: I released libtaskotron, resultsdb_api, resultsdb_frontend and execdb

Pushed libtaskotron, resultsdb_api, resultsdb_api and execdb to the infrastructure-testing repo

Added alembic to the execdb role. Added the new keys to taskotron.yaml. Deployed on dev/stg. Switched checks branch to develop on dev/stg. Let's see if there are any problems. If not, I'd like to deploy on prod on Friday or Monday if there are no objections (and no problems of course).

Issues found so far: D443 and https://phab.qadevel.cloud.fedoraproject.org/rLTRN8581d1c5296465bb074a307837f34edf586ee3e7 (no patch yet). We will need to fix both to create a new release.

After fixing docs, released and built new libtasskotron-0.3.17. infra-testing repos have been updated. libtaskotron has been updated on dev clients, assuming no big issues, will update stg later today

seems to be working OK, updated stg clients

Seems to be working great so far. What I'd like to see before switching this to production is at least one failed depcheck run, so that we can be sure that the artifact logs contain all necessary details (I already verified the other checks).

We should also write a blogpost announcing the changes and showing people where to find the logs (the Log -> link) and how to access the full logs, if needed. Maybe even send it to test-announce (or devel-announce as well?).

Setting up a blog is on my TODO list for a while now. If there are no objections, I can set up my blog (put it on planet fedora) and write the blogpost there.

Please do. No objections :)

I'd imagine something like "Packagers, hear hear! Taskotron changes coming your way..." and few screenshots with marked locations of where to click to see the logs etc. Just an idea.

@mkrizek, could you please build and deploy new resultsdb_frontend as well? We need the firefox search fix. Once we deploy this to production, we will see a major increase in people using resultsdb, we can't have the search broken. Thanks.

resultsdb_frontend built and deployed on dev/stg.

Please note that the version in the resultsdb_frontend says 1.1.5 instead of 1.1.6. It's because I got burned by make. Prior to making "bump version" commit I did make mockbuild to test if it builds. Then I did the bump version commit and built a srpm. But it didn't create a new tarball (the rule is not in .PHONY) and used the one without version bump so the rpm doesn't include proper version in init.py. Not sure it's worth rebuilding.

Has anyone seen any issues? Are we ready for deploying to prod?

We're in alpha freeze right now so we'll need to wait until alpha is released before making changes to production.

I believe this has been done for a while now, closing.

Login to comment on this ticket.

Metadata