- Full name: Robert Walker
- Date of birth: 22 October 1977
- Nationality: British
- E-mail: email@example.com
- Twitter: @rjw1
- Phone: 07810 542 241
- Address: 19 Clarendon Road, Croydon, CR0 3SJ
I am a senior operations engineer with over 17 years of experience in multiple environments using a wide range of Unixes (Linux, BSD and Solaris), configuration management tools (Puppet, Chef and CFengine2) and platforms (AWS, vCloud and bare metal) while supporting applications in a variety of languages (Ruby, Perl, Python, Java and Go). I can help infrastructure teams to define a product road map both as a technical lead and as a product owner.
I have been practicing DevOps since before the term was coined, and was an organizer of DevOpsDays London 2017. I have opinions on continuous integration and continuous deployment of both applications and infrastructure. I also believe in HumanOps, and have spoken several times about how on call should not be a burden to the people involved.
- What should wake you up at night? - London Devops #30
- How GOV.UK does on call - HumanOps London May 2016
As head of the Web Operations community at GDS, I worked with the other technical heads of community to look after our community members. I provided mentorship and line management to around 20 other engineers, and made sure they were learning and developing. I empowered the community to self-organise events and to share learnings between teams, and I also helped organise several off-site events.
I participated in the government-wide Digital, Data and Technology (DDaT) effort to define a career path for people within the software engineering community. My position as the main point of contact for recruitment of Web Operations engineers within GDS helped me realise that this was a burden that needed to be shared, and so I created a rota of people to be main contact week-to-week.
After a decision was made to change how we managed platforms, I was part of the team which enabled this transition. My main roles were liaising with my community and helping to define the work that the new central teams would do along with the existing product managers and delivery managers.
While Head of Web Operations, for a time I was also the product manager for the GOV.UK infrastructure team and participated in the planning for the whole of GOV.UK. I worked with the Government PaaS team on planning the migration of GOV.UK to the government PaaS.
As a senior engineer on the GOV.UK infrastructure team, I worked with technical architects and senior developers to define the technical direction of GOV.UK. I continued to work on several large projects which provided the foundation for migrating us to a different hosting provider. I eventually became the technical lead and product owner for the team, helping to set our direction and make technical decisions. I also helped to mentor our new staff members.
During this period, I was made Acting Head of Web Operations, at which point I relinquished my tech lead post but continued to manage the product.
I joined GDS as a member of the Infrastructure and Tools team within GOV.UK. I worked on several major projects including migrating GOV.UK to a new platform, leading technically on the setting up of a new logging infrastructure and helping with the procurement of a new CDN. I managed the team's backlog when we lacked a delivery manager, and facilitated several retrospectives. I participated fully in our second line support model and out of hours on call rota, and trained several of our developers so that they could contribute to second line support. I also helped maintain several environments and Puppet repositories.
At TagMan I was responsible for maintaining and growing a platform serving approximately two billion hits a month spread over six points of presence across the world. During my time at TagMan I worked both in the Operations Team and in a development team focused on improving the scalability of our architecture.
As part of the Operations Team, I introduced the concept of configuration management with an initial install of Puppet, and later supported a switch to Chef by writing cookbooks and managing the contractor who did the bulk of the work. I also acted as team leader during periods when we had no Operations Director.
My scalability work included identification of the inherently non-scalable parts of the codebase, helping the developers design replacement code with an eye to scalability, and liaising with product management during addition and prioritisation of the necessary tasks in the backlog. I replaced our legacy system of client configuration (which used a MySQL replication chain) with a more scalable solution based on Git and message queues. I also replaced our in-house deployment tool with a system of packages and Chef cookbooks, both to improve scalability and to bring our procedures closer to industry standards.
March 2008 – October 2010: Senior Systems Administrator for Ominor
At Ominor I set up and maintained over 30 servers for a range of e-commerce websites developed in PHP and Ruby on Rails, spread across two data centres in several environments. One of my major tasks was to provision a new rack for a client with a turnover of approximately one million pounds a month. I managed the servers with Puppet, and used Nagios and Munin to monitor the servers and collect usage data.
Throughout my time at Ominor I worked closely with the developers as part of a team following an Agile methodology. As part of this I helped maintain the continuous integration servers, and performed deploys as and when they were signed off by the clients.
January 2008 – March 2008: Contract Systems Administrator for Fotango
During the windup of Fotango, I worked as a contractor providing out-of-hours cover and daytime support while the Fotango systems were transitioned to their new provider.
January 2006 – December 2007: Senior Systems Administrator for Fotango
At Fotango I was a member of a five-person team providing support for about 200 Red Hat and Debian machines in our production, staging, development and office environments.
One of my first tasks was to design and install a rack for a new product, including redundant routers, load balancers, web servers and database servers. The routers used VRRP for failover, and the load balancers used a combination of Heartbeat, ldirectord and IPVS to provide high-availability web services. The database cluster ran PostgreSQL, and used DRBD and Heartbeat to provide a failover solution. I also set up monitoring and logging for this product using Nagios and syslog-ng. The web servers were organized as Xen guest domains spread across a cluster of machines.
I was also responsible, along with a co-worker, for investigating alternative technologies for automatic provisioning of machines, to replace the existing system based on CFEngine and a custom autoinstaller. Much of our testing involved virtualization using Xen and VMware. Our eventual choice was a combination of Puppet and FAI.
As well as projects such as these, I was part of the on-call rota providing support for our legacy systems, including our IBM SAN and several large MySQL databases.
While at Glu Mobile I installed and maintained over 20 internal and external systems which provided numerous services. My main duty was the administration of six Sun machines (3x280R, 3xv100) running Solaris which acted as a provisioning system for mobile games using Jboss and MySQL. I also administered machines running other services such as mail, DNS, Samba for internal file shares, Checkpoint FW-1/VPN-1, HTTP (both barebones Apache/PHP and Jboss/Jetty solutions) and DHCP. I installed and administered a new backup system using Veritas NetBackup 4.5 and a Sun L8, and initiated the process of migrating to Veritas NetBackup 6.0 with a Sun L100. For most of my time at Glu I was also responsible for desktop support.
At Profero I was responsible for the administration of several Intel-based machines running FreeBSD, OpenBSD and Suse Linux. These machines provided services such as mail, DNS and file sharing using Samba and AppleTalk. I also administered the machines that Profero's ad servers ran on. In conjunction with this I set up and maintained a firewall (StoneGate) which provided load balancing and clustering.
While contracting for Sciteb I installed a number of desktops (Linux and Windows) and set up several services on their Linux server.
August 1999 - June 2000: Systems Administrator and Software Developer for Pepper Head Design
I joined Pepper Head Design as an HTML coder and progressed to become a PHP and JSP developer leading a team of two other developers. I also liaised with our outsourced Java developers, and was responsible for support of desktops and servers running a combination of Linux, OS/2 and Windows.