Paper |
Title |
Page |
MOMAU007 |
How to Maintain Hundreds of Computers Offering Different Functionalities with Only Two System Administrators |
56 |
|
- R.A. Krempaska, A.G. Bertrand, C.E. Higgs, R. Kapeller, H. Lutz, M. Provenzano
Paul Scherrer Institut, 5232 Villigen PSI, Switzerland
|
|
|
The Controls section in PSI is responsible for the Control Systems of four Accelerators: two proton accelerators HIPA and PROSCAN, Swiss Light Source SLS and the Free Electron Laser (SwissFEL) Test Facility. On top of that, we have 18 additional SLS beamlines to control. The controls system is mainly composed of the so called Input Output Controllers (IOCs) which require a complete and complex computing infrastructure in order to boot, being developed, debugged and monitored. This infrastructure consists currently mainly of Linux computers like boot server, port server, or configuration server (called save and restore server). Overall, the constellation of computers and servers which compose the control system counts about five hundred Linux computers which can be split into 38 different configurations based on the work each of this system need to provide. For the administration of all this we do employ only two system administrators who are responsible for the installation, configuration and maintenance of those computers. This paper shows which tools are used to squash this difficult task: like Puppet (an open source Linux tool we further adapted) and many in-house developed tools offering an overview about computers, installation status and relations between the different servers / computers.
|
|
|
Slides MOMAU007 [0.384 MB]
|
|
|
Poster MOMAU007 [0.708 MB]
|
|
|
MOPMS020 |
High Intensity Proton Accelerator Controls Network Upgrade |
361 |
|
- R.A. Krempaska, A.G. Bertrand, F. Lendzian, H. Lutz
Paul Scherrer Institut, 5232 Villigen PSI, Switzerland
|
|
|
The High Intensity Proton Accelerator (HIPA) control system network is spread through about six buildings and has grown historically in an unorganized way. It consisted of about 25 network switches, 150 nodes and 20 operator consoles. The miscellaneous hardware infrastructure and the lack of the documentation and components overview could not guarantee anymore the reliability of the control system and facility operation. Therefore, a new network, based on modern network topology, PSI standard hardware with monitoring and detailed documentation and overview was needed. We would like to present the process how we successfully achieved this goal and the advantages of the clean and well documented network infrastructure.
|
|
|
Poster MOPMS020 [0.761 MB]
|
|
|