Please, we must check reliability, discriminant validity and convergence validity in second order of confirmatory factor analysis or only for latent variable of first of order confirmatory factor analysis.
In general, information about score reliability and validity is always welcome.
If you're asking, whether and which technical indicators of score quality are required for higher-order factors, then the answer depends on how it is you intend to use the scores from such factors, if at all.
If you're asking, does CFA which supports a second-order factor model, require additional work beyond the affirmation that the model adequately reproduces the observed relationships among the manifest variables or indicators, then the answer is "No." The reason is, the exploration of additional variables (for evidence of convergent or discriminant validity) goes beyond the customary end point of the CFA.
If you're asking, might some external gatekeeper (e.g., a journal reviewer, editor, a dissertation committee or member) ask for such information when presented with a discussion of a higher-order factor model, then the answer is "Yes."