Helen is correct; in the late nineteenth and early twentieth centuries, a number of fragments containing portions of Ben Sira in Hebrew were discovered in the Cairo Genizah (a storehouse for old or retired manuscripts attached to the synagogue in Cairo). These fragments were determined to be remnants of five distinct manuscripts, which were designated A, B, C, D, and E. They are extremely important for the textual history of Ben Sira and for the study of the Hebrew language, because before their emergence the book was known only in other languages such as Greek, Old Latin, and Syriac.
So, when the CDCH entry references "(Bmg, D)," it means that this Hebrew form is found in two of the Cairo Genizah manuscripts: B (where it appears in the margin rather than the main text; this is the meaning of "mg") and D. However, the entry suggests that the form may be a scribal error; the scribe may have intended to use a form of the verb that means "to take delight."
The asterisk at the beginning of the entry indicates that this word does not appear in Brown-Driver-Briggs (which was first published in 1906). The DCH and CDCH contain approximately 3300 of these "new words." Approximately one-third of these are words found in manuscripts discovered or published after the publication of Brown-Driver-Briggs (this category includes most of the Cairo Genizah manuscripts, as well as the Dead Sea Scrolls). The remaining two-thirds are words that have been proposed by modern scholars, based on research in cognate languages such as Ugaritic, Akkadian, and Arabic.