Hacker News Reader

Comment by grey-area | Hacker News Reader

So I would guess a huge overhaul, and typical Apple, is taking care of that vs. arguing or commenting on theories.

A design overhaul wouldn't require the dev centre to go down at all - they could just prepare all the assets etc on testing servers and switch them over when they are ready.

Given the normal warning given on any maintenance, and the obvious negative consequences for Apple's business of any extended outage (extended in this case being over a day or so), this is unintended, and most likely caused by a security breach. I think we can safely rule out a design overhaul unless Apple are incredibly incompetent.

NB itunesconnect is still up (the bit which deals with app upload, itunes store metadata etc), only the dev center - dealing with certs/device registration/distribution etc - is down.

View on HN

Keep in mind, this is the company which takes their whole store down when they need to add new products to it.

Did you mean when they have their WWDCs? In that case it might be a marketing reason.

They take it down during WWDC keynotes and other major announcements, yes.

But there are other instances where the store will go down outside of those periods. Sometimes it's for maintenance, but more often than not a new or updated product appears (e.g. a product line gets a speed bump across the board).

Perhaps they're also trying to pull together all the disparate things running on the dev servers?

nwh's comment ( https://news.ycombinator.com/item?id=6078854 ) says they've got a lot of different systems running, maybe they're trying to unify them to be on one system, making upgrading/changing easier in the future?

It's possible this is some sort of migration gone wrong I suppose, there are a lot of pieces involved in the dev centre, and I've seen a few changes popping up over the last few weeks. Seems odd that they'd just pull the whole thing though without warning, rather than just rolling back to the last state. With a migration they should have a good known state they can go back to quickly if necessary.

No planned migration or consolidation should take several days of interrupted service, it's just not necessary unless something goes horribly wrong.

Maybe they deleted all the copies of the both new and old developer centres so are having to rebuild it all from scratch.

I would hope a multi-billion dollar developer portal has a backup somewhere. It'd be insane if they didn't.

You'd be surprised the kind of risky setups multi-million <currency> companies will run with.

Reply on news.ycombinator.com