FreeBSD test results
This test was performed similarly to the Linux tests after the scripts had been trweaked. Unfortunately the time a d date on the host was not correct and this was not noticed until well after the tests had run. The results are being analyzed on Tuesday 2025-02-11 and the tests were performed over the previous weekend.
2025-02-11 corruption confirmed
The processes that stirred the pool and invoked syncoid terminated as desired but the loop that managed snapshots continued (with no effect other than to query snapshots.) The test was under way at 2025-02-04-1027 (local time on the host) and corruption was detected at 2025-02-04-2003. In other words it took about ten hours to produce corruption. Full pool status was:
root@vulcan:~ # zpool status send -v
pool: send
state: ONLINE
status: One or more devices has experienced an error resulting in data
corruption. Applications may be affected.
action: Restore the file in question if possible. Otherwise restore the
entire pool from backup.
see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-8A
scan: scrub repaired 0B in 01:32:28 with 0 errors on Tue Feb 4 16:46:00 2025
config:
NAME STATE READ WRITE CKSUM
send ONLINE 0 0 0
da1p1 ONLINE 0 0 0
errors: Permanent errors have been detected in the following files:
root@vulcan:~ #
There was a warning in the syncoid record that seems not to provide much information.
...
Sending incremental send/test/l0_1/l1_0/l2_3@syncoid_vulcan_2025-02-04:20:00:32-GMT00:00 ... syncoid_vulcan_2025-02-04:20:02:12-GMT00:00 (~ 24.5 M
B):
warning: cannot send 'send/test/l0_1/l1_0/l2_3@syncoid_vulcan_2025-02-04:20:02:12-GMT00:00': Invalid argument
Sending incremental send/test/l0_1/l1_1@syncoid_vulcan_2025-02-04:20:00:34-GMT00:00 ... syncoid_vulcan_2025-02-04:20:02:15-GMT00:00 (~ 22.5 MB):
...
Due to the time required to produce desired results, further testing on FreeBSD (at least hosting on a Pi 4B) is not a high priority.