Whilst we report on a lot of Docker hook-ups and tool additions, there are relatively few fully functioning tools on the market. I’m hereby marking today as the first (but almost definitely not the last) day that we’re going to report the GA availability of a Docker specific tool; Flocker 1.0 is now fully baked and ready to put into production.
Created by ClusterHQ, a data-oriented team of software developers, Flocker is an open source project which was created to bridge with one of the fundamental holes cracks in the container movement. Essentially, it provides developers with an operational toolkit for managing stateful services – for example databases and key-value stores – in containers. With modern applications constructed from both stateless and stateful microservices, Flocker offers an avenue for entire applications (including their state) to be containerized.
According to Michael Ferranti, VP of marketing for ClusterHQ, until the advent of Flocker, the big caveat with Docker was that it couldn’t handle data due to the inherent nature of the of the container architecture. Traditionally, when a container moves, its data volume stays in place, and the database spins up again on a new server without any data. Flocker enables developers to move both containers and their data between hosts. It also enables developers to migrate data between multiple containers, and also perform things like cross-datacenter migration.
Image Source: ClusterHQ
ClusterHQ believe that there is a strong demand for the services they can provide. Having recently commissioned a survey of DevOps.com readers, with respondents largely composed of system administrators, DevOps engineers, security and QA professionals, the team endeavoured to find out what the current state of container usage is, what the barriers to adoption are, and what types of containers people are using.
Overwhelmingly, amongst this audience of 285 typically early adopters, a staggering 40% reported that they’re already running containers in production. Of these, a large proportion are leapfrogging the traditional ‘lab’ phase for the technology, and porting the technology directly into the production cycle. Furthermore, 65% expect to use containers in production within the next 12 months.
For the bulk of respondents, somewhat predictably, it was security of container technology that was proving the biggest barrier to adoption. Interestingly for ClusterHQ, data management and persistence storage were also cited as among the top five factors for delaying uptake of Docker.
Almost everybody in the survey (95%) reported that data management in containers was of some degree of importance to their container strategy. A further 51% were worried about networking, and 48% also had concerns about developers skills and knowledge and persistent storage.
For ClusterHQ, this was “a really strong validation that people were looking for solutions before ‘Docker-ising’ everything.” Ferranti concludes that, based on their data, “You can argue that the skills and knowledge gap we’re seeing with Docker is really as a result of companies looking for a value partner who can help them make that transition.”
Whilst it’s come in for its share of FUD from certain corners, Docker remains far and away the leading container choice. In this survey, 92% of respondents said they had used or investigated it . Just 32% had dabbled with LXC , and just 20% had investigated rkt (formerly Rocket).
If you’d like to have a look at what Flocker can do for your containers, Flocker 1.0 is available as an Apache-licensed download at clusterhq.com/flocker.