due to strange behaviour in sd_journal_next(), some random starting point in
the journals were used instead of the specified "tail". as a result, lots of
log messages were (potentially) resend to the graylog server.
to nudge the sd_journal_next() call into the right behaviour, a workaround is
used, similar to what's outlined in the various related bugreports:
due to strange behaviour in sd_journal_next(), some random starting point in the journals were used instead of the specified "tail". as a result, lots of log messages were (potentially) resend to the graylog server.
to nudge the sd_journal_next() call into the right behaviour, a workaround is used, similar to what's outlined in the various related bugreports: