> So it looks like the new test failed initially but passed on a re-trigger?
Correct, that's why I added the sleep 5s (but the retrigger was without this sleep).
The failure was in getting the Administrator's kerberos ticket, i.e., a plain echo $password | timeout 5 kinit Administrator. The exit status 124 tells me it the kinit was killed by the "timeout 5s" prefix, so kinit was still thinking after 5s.
(...)
## DNS tests
Obtaining administrator kerberos ticket
Password for <email address hidden>:
autopkgtest [21:57:23]: test samba-ad-dc-provisioning-internal-dns: -----------------------]
samba-ad-dc-provisioning-internal-dns FAIL non-zero exit status 124
I rarely use the timeout command prefix, I just didn't want for a command like kinit to wait for user input forever at a prompt due to a mistake or some other failure. If it turns out this is what makes the test flaky in the future, I'll remove it.
> So it looks like the new test failed initially but passed on a re-trigger?
Correct, that's why I added the sleep 5s (but the retrigger was without this sleep).
The failure was in getting the Administrator's kerberos ticket, i.e., a plain echo $password | timeout 5 kinit Administrator. The exit status 124 tells me it the kinit was killed by the "timeout 5s" prefix, so kinit was still thinking after 5s.
(...) dc-provisioning -internal- dns: ------- ------- ------- --] dc-provisioning -internal- dns FAIL non-zero exit status 124
## DNS tests
Obtaining administrator kerberos ticket
Password for <email address hidden>:
autopkgtest [21:57:23]: test samba-ad-
samba-ad-
I rarely use the timeout command prefix, I just didn't want for a command like kinit to wait for user input forever at a prompt due to a mistake or some other failure. If it turns out this is what makes the test flaky in the future, I'll remove it.