Monthly Archives: June 2014

Iran : Yazd

We took our first bus trip to reach Yazd from Shiraz using the Hamsafar company. Booking a bus trip is as easy as it is cheap in Iran so this is by far the best way to get around even tho it’s a bit slow mostly due to the police controls along the road.

Yazd was a shock as it’s a small and beautiful desert town with an unique athmosphere. This city still haunts me and remains my favorite of the trip. The baazar, the covered streets and the mud walls gives you a feeling which is difficult to describe.

We stayed at the very pleasant Orient Hotel and spent one night in a caravanserai where I tried my luck and succeeded to rent a motorcycle for one day ! That was a fun and incredible experience and we’ll always remember the look on the amuzed face of the Iranian people when they realized some tourists where riding a motorcycle among them.










Yazd is of course not only about desert and features some beautiful and peaceful gardens.



Consul on Gentoo Linux

As a clustering and distributed architecture enthusiast, I’m naturally interested in software providing neat ways to coordinate any kind of state/configuration/you-name-it over a large number of machines.

My quest, as many of you I guess, were so far limited to tools like zookeeper (packaged on my overlay but with almost no echo) and doozerd (last commit nearly 6 months ago) which both cover some of the goals listed above with more or less flavors and elegance (sorry guys, JAVA is NOT elegant to me).

I recently heard about consul, a new attempt to solve some of those problems in an interesting way while providing some rich fuctionnalities so I went on giving it a try and naturally started packaging it so others can too.

WTF is consul ?

consul logo

Consul is a few months’ old project (and already available on Gentoo !) from the guys making Vagrant. I especially like its datacenter centric architecture, intuitive deployment and its DNS + HTTP API query mecanisms. This sounds promising so far !

This is a descripion taken from the Hashicorp’s blog :

Consul is a solution for service discovery and configuration. Consul is completely distributed, highly available, and scales to thousands of nodes and services across multiple datacenters.

Some concrete problems Consul solves: finding the services applications need (database, queue, mail server, etc.), configuring services with key/value information such as enabling maintenance mode for a web application, and health checking services so that unhealthy services aren’t used. These are just a handful of important problems Consul addresses.

Consul solves the problem of service discovery and configuration. Built on top of a foundation of rigorous academic research, Consul keeps your data safe and works with the largest of infrastructures. Consul embraces modern practices and is friendly to existing DevOps tooling.

app-admin/consul ?

This is a RFC and interest call about the packaging and availability of consul for Gentoo Linux.

The latest version and live ebuilds are present in my overlay so if you are interested, please tell me (here, IRC, email, whatever) and I’ll consider adding it to the portage tree.

I want to test it !

Now that would be helpful to get some feedback about the usability of the current packaging. So far the ebuild features what I think should cover a lot of use cases :

  • full build from sources
  • customizable consul agent init script with reload, telemetry and graceful stop support
  • web UI built from sources and installation for easy deployment
# layman -a ultrabug
# emerge -av consul

Hope this interests some of you folks !

uWSGI v2.0.5.1

This release is important to me (and my company) as it officially introduces a few features we developed for our needs and then contributed to uWSGI.

Special congratulations to my co-worker @btall for his first contribution and for those nice features to the metrics subsystem with many thanks as usual to @unbit for reviewing and merging them so quickly.

new features

  • graceful reload of mule processes (Credits: Paul Egan) : SIGHUP is now sent to mules instead of directly killing them, by default you have 60 seconds to react before a SIGKILL
  • –metrics-no-cores, –stats-no-cores, –stats-no-metrics : don’t calculate and process all those core related metrics (gevent anyone ?)
  • reset_after_push for metrics (Credits: Babacar Tall) : this metric attribute ensures that the metric value is reset to 0 or its hardcoded initial_value every time the metric is pushed to some external system (like carbon, or statsd)
  • new metric_set_max and metric_set_min helpers (Credits: Babacar Tall) : can be used to avoid having to call “metric_get“ when you need a metric to be set at a maximal or minimal value. Another simple use case is to use the “avg“ collector to calculate an average between some *max* and *min* set metrics. Available in C and python.

See the full changelog here, especially some interesting bugfixes.