Ceph

The test environment created here https://lab.fedeproxy.eu/ceph/ceph-backport

Once it works, it should probably become a standalone QA job that can be run manually. I'm not sure the CI has the ability to run docker containers. And in any case it would be overkill to run that every time although it is only needed when the script is modified.

@Nathan Weinberg, would you be so kind as to apply these patches to master and create a pull request with the following title and body?

scripts: ceph-backport.sh: minor changes to enable testing

It was tested to successfully open a backport pull request and update the corresponding issue in redmine using a local redmine instance and a local Gitea instance (it implements the same API as GitHub).

Fixes: https://tracker.ceph.com/issues/51222

The tests are implemented in a separate repository

Another issue will be opened to figure out where they should be hosted to ensure they are not lost.

@Nathan Weinberg ping ?

Thanks! I have no idea regarding reviewers right now but I'll think about it. I should have provided a title/description for the PR, sorry about that. Here they are:

Title: scripts: ceph-backport.sh: enable automated test
Description: No functional modifications. The changes are mostly about allowing an external test script (not included in this PR) to override variables and verify the script works as expected using a redmine and GitHub like (Gitea) self hosted instance.

hey Loic: egarding https://tracker.ceph.com/issues/51222 does the test environment run on any new PR related to ceph-backports? can we get it to post status to the PR or do we need to setup something, what would be your suggestion?

Deepika,

The tests were designed in an environment where I could experiment run against GitLab, not GitHub. In addition they depend on third party repositories which makes them too fragile. This was a useful first step to go from not having any kind of test environment to something that work.

The second step is to make the test self contained so they can be included in the Ceph main tree. This is what I did here: https://lab.fedeproxy.eu/ceph/ceph/-/merge_requests/1

You will see in the associated commit that the setup of the test environment, although self contained is not trivial. It indeed has a long history and should be cleaned up. But maybe this can be left for later since the main focus is to have tests running in the CI.

The tests will be triggered from:

https://lab.fedeproxy.eu/ceph/ceph/-/blob/wip-backports/src/test/CMakeLists.txt#L611

and should run in GitHub as well. You can try them by making a pull request from the content of this branch:

https://lab.fedeproxy.eu/ceph/ceph/-/tree/wip-backports

hey Loic,

I tried to test the script that ceph test would be triggering locally, but could not run it successfully... can you let me know if I am missing something,

+ for i in redmine postgresql-redmine
+ docker stop redmine
Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.
Error: no container with name or ID "redmine" found: no such container
+ true
+ docker rm redmine
Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.
Error: no container with name or ID "redmine" found: no such container
+ true
+ for i in redmine postgresql-redmine
+ docker stop postgresql-redmine
Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.
Error: no container with name or ID "postgresql-redmine" found: no such container
+ true
+ docker rm postgresql-redmine
Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.
Error: no container with name or ID "postgresql-redmine" found: no such container
+ true
+ docker volume rm --force redmine-volume
Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg.
Error: no volume with name "redmine-volume" found: no such volume

ran: ./src/test/run-ceph-backport-tests.sh

Could you please attach the full output ? These are the last lines only and suggests the redmine container was not successfully ran for some reason.

sure

2021/09/27 09:35:56 main.go:117:main() [F] Failed to run app with [gitea admin user create --access-token --admin --username root --password Wrobyak4 --email admin@example.com]: CreateUser: database is locked
Error: exec session exited with non-zero exit code 1: OCI runtime error
+ success=true
+ break
+ true
++ id -u
+ test 1212 '!=' 0
+ SUDO=sudo
+ : 127.0.0.1
+ for f in ${@:-teardown setup}
+ teardown
+ for i in redmine postgresql-redmine
+ podman stop redmine
Error: no container with name or ID redmine found: no such container
+ true
+ podman rm redmine
Error: no container with name or ID redmine found: no such container
+ true
+ for i in redmine postgresql-redmine
+ podman stop postgresql-redmine
Error: no container with name or ID postgresql-redmine found: no such container
+ true
+ podman rm postgresql-redmine
Error: no container with name or ID postgresql-redmine found: no such container
+ true
+ podman volume rm --force redmine-volume
Error: no volume with name "redmine-volume" found: no such volume
+ for f in prepare_environment ${@:-teardown setup}
+ prepare_environment
+ : 0.0.0.0
+ for f in prepare_environment ${@:-teardown setup}
+ teardown
+ podman stop gitea
8542a355aa3d170aefed3545328ba9bd23f5e2a0c19065fca1271fdf8f44c09c
+ podman rm -f gitea
8542a355aa3d170aefed3545328ba9bd23f5e2a0c19065fca1271fdf8f44c09c
++ id -u
+ test 1212 '!=' 0
+ SUDO=sudo
+ : 127.0.0.1
+ for f in ${@:-teardown setup}
+ teardown
+ for i in redmine postgresql-redmine
+ podman stop redmine
Error: no container with name or ID redmine found: no such container
+ true
+ podman rm redmine
Error: no container with name or ID redmine found: no such container
+ true
+ for i in redmine postgresql-redmine
+ podman stop postgresql-redmine
Error: no container with name or ID postgresql-redmine found: no such container
+ true
+ podman rm postgresql-redmine
Error: no container with name or ID postgresql-redmine found: no such container
+ true
+ podman volume rm --force redmine-volume
Error: no volume with name "redmine-volume" found: no such volume

I am not sure why gitea is not being pulled, i used podman instead, since I was having issues with docker

A failure to pull gitea is the blocker for sure. What kind of error message do you get when trying to pull the gitea image?

Just checking that I did not miss a question from you Deepika :-)

For the record (so I do not forget), pinged via IRC

@loicd full log: https://gist.github.com/ideepika/51526acbcff739de1ec6a031287e3b80

if I comment that out and use, I get:

ideepika@senta04:~/ceph/src/test$ /home/ideepika/ceph/src/script/backports/tests/gitea-helper.py setup
  File "/home/ideepika/ceph/src/script/backports/tests/gitea-helper.py", line 17
    d = p.dvcs(f"{self.tmp}/repository")

The problem is that:

+ podman volume rm --force redmine-volume
Error: no volume with name "redmine-volume" found: no such volume

Returns on error when the volume does not exist. The same command with docker instead of podman returns success.

$ docker volume rm --force nonexistent && echo SUCCESS
nonexistent
SUCCESS

Adding || true should fix it

Would you be so kind as to get the lastest version of this patch from https://lab.fedeproxy.eu/ceph/ceph/-/tree/wip-backports ?

You can see the output of a successful run for reference at https://lab.fedeproxy.eu/ceph/ceph/-/jobs/449

thanks Loic, I will try them out today

Hey Loic!

I tried using the script in centos 8 and I am not able to move past installing dependencies(since poetry req python3.8)

I don't have a Ubuntu 20.04 setup.

Since we are using docker, there we had faced the rate-limiting issue in the past: https://www.docker.com/increase-rate-limits

Can we move to use quay https://quay.io/?

Sure quay.io is fine !