This is the epitaph written on Balin's tomb in the Chamber of Mazarbul in Khazad-dûm. Gandalf translates it for the Fellowship as "Lord of Moria".
Tolkien provides a translation for uzbad as "lord" without any further comment. Many would point out that Khuzdul words don't start with a vowel, like uzbad does, so there must be something going on here. However, there are plenty that apparently do, such as Azanulbizar. Adunaic has words like this as well, such as inzil "flower", as does Arabic. So, I think it's perfectly plausible to see this in Khuzdul as well. For those words that do look to start with a vowel, there is an assumed glottal stop at the beginning of the word, so Khuzdul would technically write 'uzbad and 'Azanulbizar, for example. The glottal stop is a consonant seen in Semitic languages, and it allows Khuzdul words to look like they start with a vowel but actually have a consonant.
In his article "An Analysis of Dwarvish", Magnus Åberg makes the case that the u- here may be a conjunction (roughly meaning "and"), similar to the conjunction wə- found in Hebrew and wa- in Arabic. The main reason is that we don't see any other Khuzdul words that have the pattern uCCaC. His position is that there could possibly be some morphological reason that words like uzbad and inbar have the form VCCVC rather than CVCVC. However, I would argue that, based on other Khuzdul examples, it looks like these word forms are possible. With the example of Duban Azanulbizar being published, we can see the template for uzbad, uCCaC, has the same relation to the template for duban, CuCaC, as does inbar (iCCaC) to zirak (CiCaC). Adunaic also has words with similar forms to uzbad and inbar, such as igmil and inzil. Arabic has words of similar shape, and it's not uncommon for a glottal stop <'> to be prefixed, allowing these words shapes to be formed without any change in meaning. In other words, the added glottal stop at the beginning isn't counted as an extra consonant in the root. I think we can say that uzbad at least has the potential to be a valid form in Khuzdul.
After much discussion, Magnus and I agree that there isn't a specific need for a conjunction meaning "and" here. Fundinul is a patronymic, and thus part of Balin's name. Thus, the inscription doesn't translate literally as "son of Fundin and Lord of Moria". As I explain in the section on Fundinul, it would be better to translate this as "Balin Fundin-son", and then begin a new line or sentence: "Lord of Moria". However, the conjunctions wə-/wa- in Hebrew and Arabic don't always translate into English as the conjunction "and". Sometimes, they can simply start a new sentence or be used for explanatory, parenthetical purposes. The way Magnus describes this in his article is that the conjunction can be used like a "spoken comma", which is a good way to explain it. What we have here then might be "Balin Fundin-son <spoken comma> Lord of Moria". The phrase that follows the conjunction, "Lord of Moria" in this instance, clarifies or further describes the subjuct, which is Balin. This, to me, seems very logical and plausible since Khuzdul was inspired by Semitic languages.
If the u- is indeed an "explanatory, parenthetical conjunction", then the actual form of "lord" might be, as an example, zebad, similar to felak. If the initial syllable is unstressed, then it might be treated as a schwa/reduced vowel, and the u- prefix would cause the vowel to be dropped, resulting in uzbad. As another possibility, "lord" might be zâbad, and the composition form zäbad (probably just written as zabad). Again, the u- prefix would reduce the vowel, so u- + zäbad would become uzbad.
The question here is not whether this is a viable analysis of the phrase, but whether it's what Tolkien intended. On Parma Eldalamberon XVII: Words, Phrases and Passages, pg 47, we are given the following, very brief entry:
There's nothing about conjunctions, case, number, etc. Granted, there are many Khuzdul etymologies where Tolkien provides nothing but a one- or two-word English translation, so this doesn't prove anything. However, Khuzdul doesn't seem to have been a major focus for Tolkien, and the particulars of Semitic conjunctions doesn't seem likely to have been a concern here. As shown above, uzbad should be a valid word form in Khuzdul, and Tolkien provides nothing except a simple English translation. The example Uzbad Khazad-dûmu shows up as a separate line of an epitaph, which simply lists his title. So, while it's easy to see a conjunction being used here, not using one is a simpler explanation and fits the evidence. As such I think it's much more likely to be true to what Tolkien was thinking in this example.
From this, I would say uzbad is "lord" and is singular, nominative, composition. There is, in my estimation, no conjunction here.
Khazad-dûmu is mostly the same as Khazad-dûm above, except that it shows the -u ending. This matches with Felak-gundu, mênu, and the objective genitive in Adunaic. We can then say that this is "Khazad-dûm" as singular, accusative, definite (proper noun).
Uzbad Khazad-dûmu is thus probably an objective construct phrase, with noun-noun word order.
The Lord of the Rings, The Fellowship of the Ring, Book II, ch 4
Parma Eldalamberon XVII: Words, Phrases and Passages, pg 47