Assuming that this is a single-factor (1-way) analysis, this can happen for a number of reasons. The significance of the global test tells you that there are differences between groups. It doesn't tell you how strong those differences are, or where they are in your set-up. You might be able to gauge the strength of the difference by looking at the pseudo-F ratio. The pairwise tests could be non-significant for several reasons. It might be because there are not enough replicates in levels of the factor for the test to have adequate permutations, and thus power. It might be because although the global test leads to the conclusion that levels of the factor are not the same, there is large overlap in the multivariate space between levels of the factor so pairwise tests, with fewer samples, just can't reject the null hypothesis of no difference.