Hello,
For now it seems each developer has its own way of deploying Qserv
on the IN2P3 cluster. This is not fully efficient because only one
person at a time
can work on the cluster.
The main requirements I've heard yesterday, at Qserv meeting were:
- each developer needs to install its own Qserv version on the
cluster, without conflict with other developers
- each developer needs to run its own Qserv version on the cluster,
so multiple version of Qserv may run simultaneously (data is
read-only)
FYI, I've developed and tested a straightforward eups-based method:
http://www.slac.stanford.edu/exp/lsst/qserv/2015_06/HOW-TO/cluster-deployment.html
If needed I can improve it so that it support both requirement
above, it will be easy, and can be done in a few days/hours.
Please let me know if you agree with it.
Thanks,
Fabrice