Activities Overview of 2013

Q1 && Q2

Infrastructure changes

Deployed new virtualisation servers to replace ageing Phase C Sun servers and a Phase E IBM server. Old systems were starting to become unreliable and so two new IBM systems were deployed to provide the virtualisation service. These new machines have state-of-the-art CPUs and SSDs to provide much better performance and stability compared to old machines. With just two physical servers we are capable of supporting 16 different virtual machines.
Deployed new service machines. As happened with the virtualisation infrastructure, many of our servers were already more than 3 year old PhaseC machines that started becoming problematic. Therefore, 5 new IBM servers were installed in order to replace them.
Increased storage capacity of the SE to a total of 1600TB.
Installed the remaining pledged compute nodes purchased to a total capacity of 23 kHS06 (8 WNs).
Redistributed the compute nodes (WN) across the different racks to assure availability in case of power failure of one of the power supplies.

Software changes

In 2011 the WLCG community moved from a gLite CERN-based middleware stack to a new common middleware stack, named EMI (European Middleware Stack).

At CSCS, during this period, we have moved from the first release EMI-1 (certified as UMD-1) to a EMI-2 (certified as UMD-2) software release. This is an important movement forward, as the middleware stack is the glue of the WLCG resources.

Migration of most service nodes to Scientific Linux 6.
Move of all middleware stack to UMD-2 release.
Upgraded the Storage Element (SE) software to the last dCache Golden Release (2.2) that will be supported until April 2014. As happened in 2012 with the upgrade from dCache 1.9.5 to 1.9.12, this was a major change in the way the software is configured, and since it is not possible to rollback any of these updates in case of problems, extensive tests were performed in our preproduction environment (set back in 2011). Once we tested all possible scenarios and situations that we thought critical, we carried out the upgrade without any kind of problems.
Moved the Infiniband software stack in all service nodes from Mellanox-provided software, to standard kernel modules. This has proven to be the most efficient way to configure the Infiniband Network, since we moved from a convoluted upgrade process, to a standard 2-step upgrade procedure. Critical and security updates can be applied now very quickly as requested by the EGI Security Office.