Ships on both sides of the line

Homer has νηῦς and νηυσί (and νήεσσι; Aeolic νά̄εσσι), but also names in ναυ- and ναυσι-. There is also ναῦφι as a plural in various functions).

Od.8.111-120:
ὦρτο μὲν Ἀκρόνεώς τε καὶ Ὠκύαλος καὶ Ἐλατρεὺς
Ναυτεύς τε Πρυμνεύς τε καὶ Ἀγχίαλος καὶ Ἐρετμεὺς
Ποντεύς τε Πρῳρεύς τε, Θόων Ἀναβησίνεώς τε
Ἀμφίαλός θ’, υἱὸς Πολυνήου Τεκτονίδαο·
ἂν δὲ καὶ Εὐρύαλος, βροτολοιγῷ ἶσος Ἄρηϊ,
Ναυβολίδης, ὃς ἄριστος ἔην εἶδός τε δέμας τε
πάντων Φαιήκων μετ’ ἀμύμονα Λαοδάμαντα.
ἂν δ’ ἔσταν τρεῖς παῖδες ἀμύμονος Ἀλκινόοιο,
Λαοδάμας θ’ Ἅλιός τε καὶ ἀντίθεος Κλυτόνηος.
οἱ δ’ ἦ τοι πρῶτον μὲν ἐπειρήσαντο πόδεσσι·

Note -νηος and -νεως. There is also Ναυσίθοος (Od.6.7, 7.56, 7.62-63, and 8.565) and of course Ναυσικάα (Od.6.17, etc., 7.12, 8.457, and 8.464).

The key points are (L.R. Palmer, GL, p. 277):
(1) ναῦς and ναυσί show regular 'Osthoff' shortening (cf. Latin nāv-i-s for the long vowel), but νῆα, νηός, νηί, etc., of course, do not.
(2) νηῦς and νηυσί have eta by analogy with the rest of the paradigm. That analogy was not applied to the 'ship' personal names: 
(3) shortening in hiatus produces νέα, νεός (monosyllabic in Od.9.283), etc.