Citations in the OED show the word 'pack' radiating from the Low Countries to England, France, and beyond in connection with the wool trade. The earliest attestation of the noun is the Latinized neuter <pac> in an ordinance of Ghent from 1199: "Omne pac, quod in curru fertur, sive parvum, sive magnum, si fuerit funiculatum, debet quatuor denarios." 'Every pack, which is carried in a cart, whether small or great, if it has been bundled with string, owes four denarii [in duty].' The verb is Latinized in compound form in a contract of 1280 between the Florentine Cerchi and the Cistercian monastery of Melsa (Meaux) in Yorkshire: "Et inveniet dictus Abbas sarpellarios quolibet anno ad predictam lanam impaccandam et sumptibus suis cariandam usque Hulle." 'And the specified Abbot shall obtain coarsecloth sacks each year for packing the aforesaid wool in, and for transporting it to Hull at his own expense.' (The complete sentence is not in the OED, but is cited by A. Schaube, "Die Wollausfuhr Englands vom Jahre 1273", Vierteljahrschrift für Social- und Wirtschaftgeschichte 6:39-72, 159-185 [1908], p. 171, n.2.)
Du Cange has several examples (the earliest from 1177) of <paccare> 'to settle a debt, pay' and <paccator> 'debtor, payer' in Pyrenaean documents for the usual <pacare> and <pacator>, which of course represent classical Latin <pa:ca:re> 'to make peace, settle'. Here the geminate appears to be a regional orthographic device having nothing to do with our 'pack'. And the Isidorean gloss <pacceolum> 'sacculum' (i.e. acc. sg. 'small sack') appears to be a simple error for <pasc->, since Nonius defines <pasceolus> as 'ex aluta sacculus' (i.e. 'small leathern sack'), and cites Cato for its usage. Thus we have no direct relatives of 'pack' attested before 1199. The earliest such example in Du Cange, <paccare> 'in fasces colligare' (i.e. 'to bind (wool) into bundles'), dates to 1341. Subsequent citations of the OED and DC show the extension of the sense of 'packing' from wool to other merchandise as the centuries wear on.
The Germanic languages have both a strong masculine or neuter and a weak masculine noun, <pak> and <packe> in Middle Low German. Middle English <packe>, <pakke> (ca. 1225) and late Old Norse <pakki> (1337) are apparently borrowed from MLG <packe>. Late Middle High German <pack> evidently comes from the strong MLG <pak>; Middle Dutch also has <pak> (ca. 1300). The Dutch plural of <pak> is <pakken>. The Gmc. verb is weak: MLG/MD <pakken>, late MHG <packen>, late ON <pakka>.
Given the connection with wool and the Low German provenance, it is plausible to regard 'pack' as borrowed from a Nordwestblock word for 'fleece, wool', related to Greek <pókos> 'uncombed wool, fleece; tuft of wool; sheep-shearing', from the Indo-European root *pek^- (IEW 797). This same noun, IE *pók^os, became Gmc. *fahaz and is reflected as ON <fæ´r> 'sheep'. One might expect NWB *pakas. However, since the /a/ of the West Gmc. forms has not undergone /j/-umlaut, the /kk/ cannot be derived from /j/-gemination. That is, we cannot postulate an early WGmc *pakaz leading to a Class I weak verb *pakjan, later *pakkjan, for it would have become *pekkjan in Old Saxon and Old Dutch, whence MLG/MD *pekken not <pakken>. Similarly, the nouns could not have arisen from early WGmc *pakja- and *pakjan-. If the source was NWB, the gemination very likely occurred in NWB, and a form already having *pakk- was then borrowed into WGmc. Without pretending any certainty about NWB morphology, I will guess that *pakas 'fleece, wool' had the associated adjective *pakyas 'pertaining to wool', which regularly became *pakkas, and was borrowed into early WGmc as *pakkaz. This adjective was then substantivized as 'bundle or load of wool', with a masculine or neuter noun subsumed, and this in turn produced a Class II weak verb *pakko:n, later (in Ingvaeonic) *pakko:jan, and a wk. m. noun *pakkan- 'woolly mass' vel sim. When mercantilism reached the Low Countries, all three nouns *pakkaz, -am, -an were used more or less interchangeably in the wool trade. Of course, other scenarios are possible. I am merely trying to establish such an etymology as plausible.
DGK