Disable git_clone minion jobs #237
Conversation
I agree that we don't need it but we only run the "full osado" test in lower intervals. Wouldn't it be a good test for real behaviour to keep the git_clone minion jobs? Or as alternative why do we even still call fetchneedles? I guess https://progress.opensuse.org/issues/164898 at least from the title point of view was not yet completed? I created https://progress.opensuse.org/issues/180854 "Deprecate fetchneedles as we now have git_clone minion jobs" for that |
From https://progress.opensuse.org/issues/164898
|
true true. wasn't that done nevertheless as part of another ticket then? |
Nope. At least not for jobs that don't have CASEDIR/NEEDLES_DIR |
perlpunk
force-pushed
the
disable-minion-git-clone
branch
from
6bbbaba to
bad0bce
Compare
We are using fetchneedles with --depth 1 for the initial clone. We don't need to fetch the whole osado repo in this test. As a result of this we can also reduce the retry attempts and timeout again.
perlpunk
force-pushed
the
disable-minion-git-clone
branch
from
bad0bce to
304e596
Compare
openqa_install+publish and openqa_from_bootstrap are still using that for running a distri-example test |
| @@ -35,6 +35,10 @@ sub install_from_pkgs { | |||
| assert_script_run 'systemctl enable apache2'; | |||
| assert_script_run 'systemctl restart apache2'; | |||
| } | |||
| if (get_var('FULL_OPENSUSE_TEST')) { | |||
| # avoid second git fetch after fetchneedles | |||
| assert_script_run q{if [ -e /etc/openqa/openqa.ini ]; then sed -i -e 's/#\[scm git\]/[scm git]\ngit_auto_clone = no\ngit_auto_update = no/' /etc/openqa/openqa.ini; else echo -e "[scm git]\ngit_auto_clone = no\ngit_auto_update = no" > /etc/openqa/openqa.ini.d/git.ini; fi}; | |||
There was a problem hiding this comment.
Maybe we can just assume that /etc/openqa/openqa.ini.d is supported.
There was a problem hiding this comment.
Why is this in the install_from_pkgs. I just wonder if this is the best place to inject this. Following the logic from run, I wouldnt expect this hidden here maybe.
There was a problem hiding this comment.
Why is this in the
install_from_pkgs.
That's the only place where we edit openqa.ini currently. Directly below this.
There was a problem hiding this comment.
Maybe we can just assume that
/etc/openqa/openqa.ini.dis supported.
I just used the same as below, so I'd say it would just look weird if we still have that logic below but not in the code I introduce.
There was a problem hiding this comment.
Dont want to be destructive but I think it make sense to take those openqa.ini out of that function, wrap them into another with a corresponding name and add it in the run function. it will look more clear and will keep the functions clean. wdyt?
There was a problem hiding this comment.
I'm not sure why you're all complaining about this line but did not complain about line 43 below
There was a problem hiding this comment.
Because line 43 was introduced when the new config parsing was just introduced so it had to support the old and new behavior. However, I don't think this is required anymore so always creating a drop-in config file should be enough.
We are using fetchneedles with --depth 1 for the initial clone. We don't need to fetch the whole osado repo in this test.
As a result of this we can also reduce the retry attempts and timeout again.
After #236