Commit Graph

814 Commits (4fb8548aba150dfb3c631c36702d3723e2f0b79c)
 

Author SHA1 Message Date
Christopher Usher 1d09e28b1e fixes to ekimekim's suggestions 6 years ago
Christopher Usher c81d538a79 thrimshim seems to be working 6 years ago
Christopher Usher e4fc878577 logic of the post 6 years ago
Christopher Usher 4c5b6e4cda GET working 6 years ago
Christopher Usher 5faa70dfc2 getting thrimshim to build and run is a minor success 6 years ago
Christopher Usher 57597c94cd hopefully some progress on the thrimshim 6 years ago
Christopher Usher be41be7878 Initial thrimshim commit 6 years ago
Christopher Usher 072e51f287 Renaming a variable that should have been part of the last commit 6 years ago
Christopher Usher 61107346c8 Fixed backing off on exceptions and some more documenation 6 years ago
Christopher Usher 728adb7c1d improvements suggested by ekim 6 years ago
Christopher Usher 530b9f7d5e more improvements based on ekims comments 6 years ago
Chris Usher 332e03de80 started in on ekim's comments 6 years ago
Christopher Usher 2857d3fb9f comments and some whitespace handling 6 years ago
Christopher Usher 4e6dbe1c74 Added localhost option to backfill to avoid backfilling the local machine 6 years ago
Christopher Usher ade0ad3d18 rewrite of get_nodes to allow getting list of files from a file 6 years ago
Christopher Usher 23fea7b154 bug fixing after testing 6 years ago
Christopher Usher 149974ce54 added multiple streams by largely copy and pasting the code from the
downloader
6 years ago
Christopher Usher e4364b75b1 options to change where the node list is coming from 6 years ago
Christopher Usher baae0f1ac1 bug fix in arg list 6 years ago
Christopher Usher 65143a8ca2 more flexability for start time 6 years ago
Christopher Usher a8cb1ff370 fixed start not propagating to list_hours plus some refactorting 6 years ago
Christopher Usher 57bb74632f I should test these changes soon 6 years ago
Christopher Usher 64bc76c48b error handling I guess 6 years ago
Christopher Usher 09368d92e1 fixes and improvements suggested by ekimekim
* simplied the backfiller local - now just a full backfill every couple minuteso
6 years ago
Christopher Usher 4eac6189ce backfiller working in parallel 6 years ago
Christopher Usher f4385ad4e3 hopefully did break anything with this refactor 6 years ago
Christopher Usher 1f53fa8d29 Bug fixes and logging improvements to the backfiller 6 years ago
Christopher Usher c9f6ee95c5 clean up for new gevent based backfiller. 6 years ago
Christopher Usher 7d9a5b4626 added workers and a worker manager 6 years ago
Christopher Usher be8d40d1ba Move the code for calculating hours outside the code that backfills 6 years ago
Chris Usher ed58b6e44d reintroduced a start time for the backfiller; more logging 6 years ago
Mike Lang 292188ad7c database: Remove retry_on_conflict helper and default to autocommit
All our usage was of a single query anyway, so autocommit is easier to handle.
You can still opt into a longer transaction using the transaction() helper.
6 years ago
Mike Lang 73640ed4ab database: Add column video_id for storing upload-location-specific metadata for identifying video
ie. for youtube, the video id.
6 years ago
Mike Lang dc2eb6ed74 Add some common database code
This code manages the database connections, setting their isolation level correctly
and ensuring the idempotent schema is applied before they're used.

Applying the schema on startup means we don't need to deal with the database's state,
setting it up before running, running migrations etc. However, it does put constraints on
the changes we can safely make.

Our use of seralizable isolation means that all transactions can be treated as fully
independent - the server must behave as though they'd been run seperately in some valid order.
This will give us the least surprising results when multiple connections try to modify the same
data, though we'll need to deal with occasional transaction commit failures due to conficts.
6 years ago
Mike Lang cea66a4bbf database: Rename start/end to event_start/end, add channel and quality
* Had to rename `end` as `end` is a reserved word in postgres SQL.
`event_end` is more consistent with `video_end` anyway. Updated `start` to match.

* Added ability to specify channel and stream quality in the editor, which may prove useful
if we have issues with a particular stream quality, or if content needs to be captured from
other channels.
6 years ago
MasterGunner 7423f8c4ef Update DATABASE.md
Changed upload_location to be edit input.
6 years ago
MasterGunner d89458c27d Update DATABASE.md
Changed allow_holes and uploader_whitelist to be edit inputs - there's no need for them to come from the sheet; and we'll have an admin dashboard for modifying them if needed.
6 years ago
Mike Lang 437d38e646 DATABASE.md: Add image_links column
This solves the problem of rows which don't need a full cut video,
but we'd like to link to an image or a short gif or clip of it.
It is a sheet input that is only used in the output sheet, so it doesn't affect the wubloader itself.
6 years ago
Mike Lang df66553b38 downloader: Start backdoor later so workers is in locals 6 years ago
Mike Lang 86da9d9fe8 downloader: Support watching multiple channels
This is useful eg. for watching db_admin or other testing channels in addition to the main channel.
6 years ago
Mike Lang f0d9aa82c2 Ignore segments that are marked as ads
* Checks for the SCTE35-OUT/SCTE35-IN marks in the HLS stream that indicate an ad start/end
* Ignores those segments completely
* Doesn't mark the StreamWorker as up until it sees the first non-ad segment

Some other operational notes:
* The main risk this adds is that re-connecting / refreshing master playlist takes longer.
  If all downloaders are doing this at the same time (ie. because the stream only just came up,
  or during a deployment rollout), all downloaders might be waiting for ads to finish and
  you'll miss segments.
* We should run more downloaders to compensate. This also increases the chance at least one of
  them won't get any ads, so we get everything right from stream-up.
* The other mitigation we can do is have geographically diverse downloaders. This decreases the risk
  that they all get served an ad, and at least at time of writing it seems that no in-stream ads
  are served outside of these regions:

> US, Canada, Germany, France, Sweden, Belgium, Poland, Norway, Finland, Denmark, Netherlands, Italy, Spain, Switzerland, Austria, Portugal, UK, Australia, New Zealand
6 years ago
Mike Lang ef0a78fce3 Updates to database states and columns
Split UPLOADED into TRANSCODING and DONE, to represent the time after upload
that youtube is transcoding the video and it's not viewable.

Any cutter can poll for the state of a transcoding video and mark it as done.

Add some extra sheet input columns.
6 years ago
Mike Lang 6b3a0fea9f Add a doc covering how the database is used
I fully expect the exact list of sheet inputs, edit inputs and outputs to change.
The important thing I wanted to codify here was the state machine and the behaviour of the cutters.
6 years ago
Mike Lang c0f94059aa downloader: Stop retrying in SegmentGetter after a long timeout
In resource contention scenarios, all calls can start failing due to
not being able to read the response in a timely manner.
This means SegmentGetters never stop retrying, leading to further contention
and a feedback loop.
We attempt to put at least some cap on this scenario by giving up
if an amount of time has elapsed to the point that we know our URL couldn't be valid anymore.
Since we don't actually know how long segment URLs are valid, we are very conservative about
this time, for now setting it to 20min.
6 years ago
Mike Lang 81aee0ee1e Increase hard timeout for getting segment headers
When we're under CPU or disk contention, doing other work
can become very slow. We want to avoid spurious errors in this situation
as this causes further retries and further contention.

One easy way to do this is to increase the time we have to finish fetching headers.
6 years ago
Mike Lang 787b9002ab restreamer: Use correct name for dateutil 6 years ago
Mike Lang 3a1e4b0aef restreamer: Fix missing dependency
This was hidden because common included it
6 years ago
Mike Lang 997c1242b2 get_best_segments: Let other things run
get_best_segments can sometimes take a very long time,
we don't want to stop other work from happening while it's ongoing.
So we ask gevent to run other things until there's no other work to do,
then we do one hour, then check back with gevent again.

In combination with the performance improvements, this should mean we don't block
other things from running for more than a few hundred ms at most.
6 years ago
Mike Lang bf08aa29b8 parse_segment_path: Use datetime.strptime instead of dateutil.parser
strptime is much faster but can't handle as varied formats.
But in this case we fully control the format, so there's no reason not to use it.

Profiling suggests we spend about 80% of our time in get_best_segments just parsing dates,
so this is a signifigant performance gain.
6 years ago
Mike Lang bcdb268ce8 Also need to replace locks on the counter float values to prevent deadlocks
See comment for full details
6 years ago