Overview

Following the BigData & Cloud Design Insight for our platform requirements (ie. cheap hardware, horizontal scalability, fault tolerance...), the Cersat platform is quite basic on hardware point of view. There are 3 different hardware components :

• 100 x servers for both storage and processing
• 20 x servers for specific services (main schedulers and masters, VmWare ESX, other identified applications ...)
• 6 x network switches (with a lot of spaghetti links ;-) )

TODO: faire article ! Indiquer avantages/inconvenient, type de monitoring (hard, services, processings), alert/etat courant, graphes etc...

Managing such a cluster requires some tools to raise alerts when anomalies occur (hardware failures or system issues) but also to investigate current or past state of the cluster. Here are a few tools we use at Cersat.

Current State

OpenStack platform is no more used nor maintained at Cersat. We consider this solution really heavy considering our simple needs, and other more simpler solutions were more appropriate in our case.

We switched in 20xx to a homemade solution based on the well known "chroot" command. In a ...

TODO : ARTICLE !

This article details how we manage daily the cluster, from the system administration point of view. Remote management for a hundred nodes, without strong homogeneity guarantee, can be complicated. Here are some tools and practices we use at Cersat.

Unless you have a scalable desktop ;-):

TODO : continuer cet article : cssh ...

Design insight

This project started in 2010, main plaform design concepts were then coming from emerging Big Data and Cloud Infrastructure-As-A-Service (IAAS) technologies.

Why using IAAS / PAAS at Cersat ?

Infrastructure and Platform as a Service (IAAS, PAAS) are useful feature for a satellite data center like the Cersat : providing a multi-OS environment following projects requirements allows a lot of flexibility compared to constraining projects to use existing environments or deploy specific hardware environments ...

Purposes

At Cersat, we manage and use everyday data from dozens of satellite mission archives. This requires a big storage archive capacity since satellite datasets are nowadays about a few hundred gigabyte to hundreds of terabytes, and even petabyte for new missions.

Introduction

We are now using MooseFS distributed filesystem for more than 3 years in production. A frequent question is : how is your feedback managing a petascale distributed filesystem using MooseFS ? This article aims to answer (objectively if possible) to this question.

Service nodes

Compute nodes

Network

Overview

The platform software architecture relies on several layers, each containing one or more components :

• Distributed Processing : Torque/Maui, Hadoop Map/Reduce, Celery, Gearman
• Virtual Environments / Containers : Chroot, Docker, Eucalyptus/OpenStack
• Distributed Filesystem : MooseFS, Hadoop HDFS
• Bare Metal / OS : Ubuntu 12.04 LTS

La plateforme software Cersat est composée de :

• MooseFS
• Docker
• Torque/Maui
• Hadoop

TODO : ARTICLE ! + photos du tout debut, avec les cerhouse etc..