Add a test case that creates and starts two nodes, has one assume the
primary role and the other the secondary role, writes to the device,
verifies that the data is synced to both nodes, then switches the roles,
and finally shuts both nodes down and compares their backing stores.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

des created this revision.Sep 14 2026, 10:15 PM

Herald added a subscriber: imp. · View Herald TranscriptSep 14 2026, 10:15 PM

des requested review of this revision.Sep 14 2026, 10:15 PM

Harbormaster completed remote builds in B76897: Diff 186727.Sep 14 2026, 10:15 PM

des added a parent revision: D59675: hastd: Allow overriding the node name.Sep 14 2026, 10:15 PM

des added a reviewer: tests.Sep 15 2026, 3:39 PM

This relies specifically on IPv4. Can you use localhost instead of 127.0.0.1

sbin/hastd/tests/hastd_test.sh
143–146	This can be handled in the cleanup routine instead.
156–161	This just handles the cleanup once instead of doing it inline in the test.

des marked 2 inline comments as done.Sep 25 2026, 10:56 AM

des added inline comments.

sbin/hastd/tests/hastd_test.sh
143–146	Then we don't test that `hastctl` is capable of stopping `hastd`.

ngie added a subscriber: jmmv.Sep 26 2026, 12:07 PM

ngie added inline comments.

sbin/hastd/tests/hastd_test.sh
12	BTW: I'd give a ship-it for this change right now -- thank you :)!
143–146	Removing the files with `rm -f` (and without `atf_check`) masks issues if the files are missing. That seems less than ideal for the tests if your goal was to confirm that the services are no longer running if the pidfiles are gone. At the very least leaving the files behind and cleaning them up in cleanup with emit warnings when kyua completes the test -- @jmmv filed an issue for no longer masking cleanup problems at test exit. It might be worth investing in that for a later release [if it's tunable via the config] so items like this can be caught and asserted on loudly.

des marked an inline comment as done.Sep 29 2026, 12:19 PM

des added inline comments.

sbin/hastd/tests/hastd_test.sh
143–146	Stopping the service does not remove the pidfile.

ngie added inline comments.Oct 1 2026, 5:58 AM

sbin/hastd/tests/hastd_test.sh
143–146	Stopping the service does not remove the pidfile. Really? That seems like a bug in and of itself--especially since the service uses the pidfile(3) APIs to manage its PID files.

des added inline comments.Oct 1 2026, 7:26 AM

sbin/hastd/tests/hastd_test.sh
143–146	It has never been a requirement. Looking at the code, it appears to remove its pidfile only when it fails to start, not when it terminates after having successfully started.