philipcristiano

I’m using forgejo mirrors to push my local repo to GitHub, seems to work well.

https://forgejo.org/docs/latest/user/repo-mirror/

Depending on the version of the iPhone, they had disadvantages such as not being able to send MMS.

https://www.imore.com/1st-generation-iphone-owners-mms

When I tried this I was hoping there was some reduction in packaging associated. At least for me, it was just all the same boxes on the same day, instead of one big box, or something reusable/returnable like Fresh.

What would you use to provide a single endpoint to multiple instances of an app server?

The article claims "Without its internal heaters turned up to-the-max, a shrew would leak out all of its heat and freeze. And this explains its furious appetite."

Dairy and eggs both involve the death of animals. Dairy cows produce milk for their offspring. In order for us to get milk the offspring must be removed. Typically males are sold for veal, females continue as dairy cows. The production of laying hens results in roughly 50/50 male vs female chickens. Most of the males are killed shortly after hatching as they are not the variety kept for meat.

The autoscaler will add minions when there is a pod that cannot be scheduled. It doesn't help in balancing existing things, but at least the new pod should be able to go there.

kops creates an autoscaling group for the worker nodes as part of the cluster creation process. Adding the cluster autoscaler is simple as deploying the autoscaler (as mentioned in your linked docs) pointing to the correct ASG. The IAM permissions for the autoscaler can be added with `kops edit cluster $CLUSTER`.

What is the concern with using bigger nodes then planned?

I agree the basic networking has a lot of limitations. Compared with added more layers with networking, I'd rather have a simpler setup with fewer nodes, even if they are larger.

Pentagon heavily relies on kops for managing Kubernetes. There is an open issue for kops to support DO. https://github.com/kubernetes/kops/issues/2150.

CockroachDB is inspired by Spanner and available now. Github page says "CockroachDB is production-ready"

https://github.com/cockroachdb/cockroach

Cows can live that long but after production drops (about 4 years) they will be sold off for beef. Not breeding cows would definitely help.

https://en.wikipedia.org/wiki/Dairy_cattle lists a few sources for the 4 years figure.

That's the case if you were going to deploy this on Linux. The packages provided are for SmartOS.

I'm surprised you found that, I'm working on etsdb on hack days when I get the chance. It sort of works but shouldn't be trusted with production data.

At least with that page I "did not receive X-UIDH header" on T-Mobile. Didn't check for headers with other names.

Thanks for that link! Definitely a case (for me) where the news around the original outrage reached more than the truth.

It will be interesting to see the uptake of the program given the many claims for where Makerbot has taken community designs and applied for patents. Last article I remembering seeing was quite recent (http://traverseda.wordpress.com/2014/05/23/makerbot-blatentl...)

Can you explain the reason for this though? It seems like the site is talking about the actual website Tumblr, I'm not confused about the brand in question. I'm not well informed from a trademark stance but it seems like as a consumer (and this response page) this site is not confusing the trademark at all. Was it previously using logos/wording saying that it was Tumblr?

It doesn't accept, not sure what the queue length is, either 100 or 1k as a guess.

That's definitely an issue and you have to weight the time it takes to work around that with how long you mind the service being unreachable.

A new AJAX endpoint is pretty simple to work around with 2 releases, one to add the endpoint and another to use it. Most changes probably aren't that fortunate.

While not great for migration uWSGI also has chain reloading [1] which will do the same thing as killing each worker. I use Tornado at work for some applications so I had to come up with something for Gunicorn as uWSGI + Tornado doesn't really fit together.

  [1] http://uwsgi-docs.readthedocs.org/en/latest/Changelog-1.9.html#chain-reloading

The issue with that is that it tends to stop all the workers then start them again, not one-by-one. Since some applications can take a minute to start that leaves the socket open but won't start accepting connections until the workers start again.

That's going to be the next step for these as requests come in for images of low aspect ratio which generally just become resized smaller instead of cropped around a face.

We haven't trained our model at all and just use the include XML file. We have adjusted the default neighbors threshold which helps mitigate the knee factor. Faces tend to be in the 40+ neighbors range from my trials and the default is 3.

We use ImageMagick for cropping and resizing, PIL was just easier for the example. I haven't noticed any issue with PIL but I also haven't used it for any serious image processing.

Beards and sunglasses also tend to cause problems. It's also nice to allow manual setting of where faces in some web editor to override what OpenCV might find.

Thanks for posting that. I haven't seen that one in a while but built a few. I'd recommend if anyone uses it to not expose it directly to the internet, once someone malicious sees all the sizes in the URL they may try hitting 1..n X 1..M which is a big hard on the servers.

We used specifically named sizes to avoid that problem although whitelisting in the edge servers would also work.

Were you using your own test data or a provided XML file?

Sports shots in our case don't get a great hit rate. Adjusting the `minNeighbors` parameter can help out with that depending on how many false positives you can accept. Musical artist misses are in the single digit percentages although shadows and strange backgrounds can give some additional faces that we don't really care for.

When collecting images we are now searching for those with more direct faces visible to make the detection easier. After that though we just try to get the face in the direct center and fall back to hoping the face is in that spot if we can't detect any.

At some point I want to try checking for partial face matches as well which should help in major sports since we tend not to use headshots.

I was planning to do something similar after I learned about Mesos although I haven't found the time. At the moment I'm using a few processes with MySQL publishing with RabbitMQ to handle distribution and node failures. It doesn't have nearly as nice an interface but RabbitMQ isn't terribly difficult to manage. Thanks for open sourcing!

We use StatsD at SeatGeek (any at my previous job as well) to track as much as possible. In general we try to time each call to an external service and use counters for any exceptions with those services.

As someone who lives just outside Philadelphia I would disagree. At least the people I know who have left the city do so because they only came here for school and take jobs back home.