das: Retry getting header in catch-up routine a few times before failing out

[renaynay]

Retry getting header at given height inside catchUp a few times (maybe 3?) before failing out entirely.

Alternative would be to retry the header a few times before giving up. But could be done in a separate PR.

Originally posted by @liamsi in #473 (comment)

[Wondertan]

For context:

[Wondertan]

But, there is no point to retry getting headers stored locally. We reliably know if we have it or if don't.

Even in the case where we fetch headers from the network via HeaderExchange, it should retry on its own, not the user(DASer) of should implement the retry. So I am closing this issue, see #497 which mentions retries for HeaderExchange.

In general, I doubt the idea of requesting single headers by height and not by hash(as we do here), because we cannot verify them, so even the idea of requesting headers here over the net does not seem to be valid.

This is not the case if we would stick to backward syncing/always identify headers by hash approach