"digdelv" system test often fails the "try the next server after a TCP socket connection error/timeout" check
For at least the past two weeks or so, the following check in the
digdelv
system test has been failing particularly often, for no
apparent reason that I could think of:
2024-01-14 04:17:49 INFO:digdelv I:digdelv_tmp_aq8itlsn:check that dig tries the next server after a TCP socket connection error/timeout (89)
2024-01-14 04:18:09 INFO:digdelv I:digdelv_tmp_aq8itlsn:failed
It is an intermittent failure. It seems to be at least strongly "preferring" FreeBSD jobs (and may even be exclusive to these, but I have not checked all of its occurrences):
- https://gitlab.isc.org/isc-projects/bind9/-/jobs/3926964
- https://gitlab.isc.org/isc-private/bind9/-/jobs/3938693
- https://gitlab.isc.org/isc-private/bind9/-/jobs/3938692
- https://gitlab.isc.org/isc-private/bind9/-/jobs/3939505
- https://gitlab.isc.org/isc-private/bind9/-/jobs/3939504
- https://gitlab.isc.org/isc-private/bind9/-/jobs/3939503
The check itself dates back to March 2022 (!5976 (merged)), so it is surprising
to only see an uptick in its failures now. Perhaps some other fix
merged in the meantime changed dig
behavior in a way that trips this
logic up?