Skip to main content

How To Fail At Upload APIs

Youtube, what the hell is up with your upload APIs? Here we are, hacking along and being all "Hey, we got accounts syncing and thumbnails popping up and videos getting attached to blog posts and its nifty-pie, oh yeah." when we make the move to be even more web 2.0y and add full authentication with youtube accounts and integrate the full upload cycle into the media selection. We take some little video upload tests and we're happy about that. Oh, that was a nice milestone to hit in FogBugz, I tell you what.

The week passes and suddenly I'm asked why all the video uploads keep failing. All the usual things are checked, but the upload tokens validate and the headers are correct. All our tests are within the upload limits, too. These aren't friendly "Hey, I really hate to tell you this, but the video you just uploaded didn't go so well. Can you try it again in a bit? Thanks"-errors, either. These are "Fuck off. I just reset your HTTP connection, bitch"-errors. Other times, we get 502 Bad Gateway from Youtube's servers and that isn't the kind of nice error you expect from a professional service like that. With the errors happening on the client browser, I'm left with nothing to do by way of responding nicely. Our own machines never get a single byte from Youtube on the matter, much less the nicely formatted response-requests they promise to use to tell us all about the sucess or failure of the uploads we set our users up with.

What gives with the weird error and what gives with they not being nicer about it?

Through support forums our problem is matched up not with some obscure thing, but to many, many posts. It seems like everyone and their dog gets reset connections and 502 errors during uploads. At this point, I'm absolutely questioning that Youtube was ready to release this part of the API when they did, because it is obviously not mature. Now, I happen to know it on good authority that some people do upload videos to Youtube, from time to time. You know, with that little uploader they wrote called The Youtube Site Itself. So, the theories I have is that they either aren't eating their own dogfood or their own uploader is doing smart-ish upload resuming when their internal upload API chokes on them constantly. In the end, I have to ask, "What gives?"

Apparently, according to support responses and our own tests, if you upload to uploads2.gdata.youtube.com instead of uploads.gdata.youtube.com, it works fine. This is their "new upload system", but what does that mean? The API is completely identical, so does "new upload system" mean a new specific set of boxes that handle uploads? Specific boxes doesn't seem very cloud-like for Google, don't you agree? Well, i

Comments

Popular posts from this blog

CARDIAC: The Cardboard Computer

I am just so excited about this. CARDIAC. The Cardboard Computer. How cool is that? This piece of history is amazing and better than that: it is extremely accessible. This fantastic design was built in 1969 by David Hagelbarger at Bell Labs to explain what computers were to those who would otherwise have no exposure to them. Miraculously, the CARDIAC (CARDboard Interactive Aid to Computation) was able to actually function as a slow and rudimentary computer.  One of the most fascinating aspects of this gem is that at the time of its publication the scope it was able to demonstrate was actually useful in explaining what a computer was. Could you imagine trying to explain computers today with anything close to the CARDIAC? It had 100 memory locations and only ten instructions. The memory held signed 3-digit numbers (-999 through 999) and instructions could be encoded such that the first digit was the instruction and the second two digits were the address of memory to operate on

Statement Functions

At a small suggestion in #python, I wrote up a simple module that allows the use of many python statements in places requiring statements. This post serves as the announcement and documentation. You can find the release here . The pattern is the statement's keyword appended with a single underscore, so the first, of course, is print_. The example writes 'some+text' to an IOString for a URL query string. This mostly follows what it seems the print function will be in py3k. print_("some", "text", outfile=query_iostring, sep="+", end="") An obvious second choice was to wrap if statements. They take a condition value, and expect a truth value or callback an an optional else value or callback. Values and callbacks are named if_true, cb_true, if_false, and cb_false. if_(raw_input("Continue?")=="Y", cb_true=play_game, cb_false=quit) Of course, often your else might be an error case, so raising an exception could be u

How To use Sphinx Autodoc on ReadTheDocs with a Django application

Sphinx is awesome for writing documentation. ReadTheDocs is awesome for hosting it. Autodocs are great for covering your entire API easily. Django is a great framework that makes my job easier. Between these four things is an interaction that only brought me pain, however. I'm here to help the next dev avoid this. Autodocs works by importing your modules and walking over the classes and functions to build documentation out of the existing docstrings. It can be used to generate complete API docs quickly and keep them in sync with the libraries existing docstrings, so you won't get conflicts between your docs and your code. Fantastic. This creates a problem when used with Django applications, where many things cannot be imported unless a valid settings module can be found. This can prevent a hurdle in some situations, and requires a little boilerplate to get working properly with Sphinx. It require a little extra to get working on ReadTheDocs. What makes this particularly h