Ironically, the talk is about Autoscaling
And Google Locations is on its way out
So Precise Pangolin was published as Ubuntu 12.04 LTS on April 26, 2012.
That’s a long time ago. Back then, Battleship, The Avengers (3D) and Cabin In The Woods (3D) were released. Intel released the Ivy Bridge Microarchitecture. The last proper US president campaigned for his second term and the US weren’t a failed state back then. It was a different world.
A blog post over at Percona discusses better replication for MySQL and compares Galera and MySQL Group Replication.
Galera builds their own initial state transfer mechanism and their own transaction distribution mechanism, independently of MySQL replication (write set replication wsrep). wsrep is synchronous – on commit, the write set is shipped, applied and acknowledged (or not).
MySQL Group Replication strives to achive the same thing, but uses their own, “MySQL native” set of technologies to do this.
This is a replay of a much older blog post, which was available in German language in the old blog. It’s from 2012, and neither GTID nor Galera cluster or Group Replication existed back then.
Wonka> The http://www.toppoint.de probably will never have meaningful load, but I would like to know how one would make this highly available. Some kind of Redundant Array of Inexpensive Databases.
Lalufu> MySQL with replication? Or DRBD?
Isotopp> With DRBD. Not with replication.
The Original Mac Mini G4 had a single-core 32-Bit PowerPC CPU at some 1.25 GHz with 512 KB on-chip Cache. It was released in 2005.
It has a Geekbench Score of 766.
The Raspberry Pi basically is a a cellphone, without the battery and the packaging. The Raspberry Pi 3 was released in 2016, and has a BCM 2837 with a quadcore ARM with 1.2 GHz.
It has a Geekbench Score of 2128.
In case you need to illustrate how Moore’s Law works, in a practical, touchable way.
Erik Bernhardsson has been running Big Data on Git repositories of various kinds.
He was trying to find out what the half-life of code is. That is, when you commit to a repository, your code becomes part of a project, but eventually other code will replace it and it will no longer be part of the current version. How stable is the codebase, what is the half-life of code? And why is it different in different projects?
As a project evolves, does the new code just add on top of the old code? Or does it replace the old code slowly over time? In order to understand this, I built a little thing to analyze Git projects, with help from the formidable GitPython project. The idea is to go back in history historical and run a git blame […]
Somebody sent me a link to Vice withe the comment “A multiple hit in the Venn Diagram of your interests”.
It’s about an artist using technology disguised as ritual magic to trap self-driving cars (and similar shenanigans). The assessent was correct, this is beautiful.
The image from the article shown above shows a self-driving car inside fake street markings. The broken lines allow the cars logic to enter the circle, the unbroken linkes mark a demarcation that must not be crossed, hence the car can never leave.
It ties back to a story my driving instructor told me. He was making a point about “How things are being presented matters”, relating about a beginners driver who had been told to imagine unbroken lines as a “wall that cannot be crossed” and who because of that had problems – sometimes rules must be broken to preserve their meaning and spirit.
I will be speaking at the Netways Open Source Data Center Conference, which is in Berlin between May 16 and 18.
At work, we are currently busy loading our first two Kubernetes Clusters (Openshift actually) with workloads.
What exactly will be in the slides I do not know, yet, but it will be about our journey at Booking, the transition from automated baremetal provisioning of rather monolithic applications to a more containerized setup and the changes and challenges this brings. It will be very much a snapshot of the state of things at that point in time, and our learnings and perspective then.
Since 5.0, MySQL does allow natively encrypted connections to the database, and supposedly also does support client certs for user authentication. Supposedly, because I never tried.
MySQL as a database performs well with transient connections as they are prevalent in two-tier deployments (mod_php, mod_perl, mod_python to database), in which a database connection is made upon web request, and the connection is torn down at the end of the request. This model does not scale so well with encryption in the mix, as on connection a full TLS/SSL exchange must be made.