Closed pdmthorsrud closed 4 years ago
I do not think that github-actions runner will be able to run this pipeline, due to resource limitation
I'll fully admit I forgot about that. Did learn about actions though! Luckily there was very little time spent on the PR itself.
It would be any way useful to have target test
in Makefile, be able to run tests locally
It would be any way useful to have target
test
in Makefile, be able to run tests locally
~I agree with that.~ Considering terraform runs in attached state an ansible task checking the health of all services is redundant. The terraform-task would never exit without everything being healthy.
Okay, so miraculously the tests passed. It says black on white that the machines have 7gb RAM available ref.
~This seems like a fluke, but I'm rerunning to see if it's reproducable.~ Ran successfully three times in a row, hovering around 10-12 min execution time, which is expected. @fredrikhgrelland do you have an opinion on this?
It would be any way useful to have target
test
in Makefile, be able to run tests locally~I agree with that.~ Considering terraform runs in attached state an ansible task checking the health of all services is redundant. The terraform-task would never exit without everything being healthy.
Are you sure about this? Can you confirm that we have all services passed?
Strange.. It sure reports two cores and two threads https://github.com/fredrikhgrelland/data-mesh/pull/91/checks?check_run_id=780740006#step:3:1120 on "mac os level" but the hardware is 4 cores: https://github.com/fredrikhgrelland/data-mesh/pull/91/checks?check_run_id=780740006#step:3:1009 and memory is 12GB ? https://github.com/fredrikhgrelland/data-mesh/pull/91/checks?check_run_id=780740006#step:3:1007
This has to be the actual hardware: https://everymac.com/systems/apple/mac_pro/specs/mac-pro-six-core-3.5-xeon-e5-gray-black-cylinder-late-2013-specs.html
Plenty storage space as well: https://github.com/fredrikhgrelland/data-mesh/pull/91/checks?check_run_id=780826831#step:4:51165
Say what now... nice detective work, haha. I guess we might as well keep the tests for now then, and remove it later if it for some reason starts failing.
Are you sure about this? Can you confirm that we have all services passed?
That answer was based on a comment from you, which was that the nomad_job
resource would run until the service is healthy in nomad, and by extension, consul.
After a little digging: ref, if detach = false
, the provider will keep running until the deployment is successful, or it fails and returns an error.
I do not think that github-actions runner will be able to run this pipeline, due to resource limitation